refactor: gateway request record and capabilities (ADR-0002:S7, S8) (#1363 )

* test: pin gateway pipeline order, request ids and region * refactor: carry request facts in one ingress request record * fix: seal the request principal once * refactor: build the caller ARN once in credverify * refactor: gateway takes policy and admission as capabilities * refactor: gateway takes cluster readiness as a capability * chore: gofmt GatewayConfig field alignment * docs: record slice 11a request record and capabilities * docs: add PR number to slice 11a migration row * refactor: move pipeline capability interfaces to ingress/aws/pipeline
ci: ask PR authors whether Mulga design needs to change (#1362 )
2026-10-31 16:34:49 +00:00 · 2026-10-31 17:57:15 +10:00 · 2026-10-31 15:56:46 +10:00 · 2026-10-30 23:12:58 -05:00 · 2026-10-31 14:07:05 +10:00 · 2026-10-30 22:53:45 -05:00
2107 changed files with 128255 additions and 36510 deletions
@@ -0,0 +1,3 @@
+# The committed bundle carries the AWS SDK's template strings verbatim,
+# including trailing spaces that are part of their values.
+spinifex/runtime/roles/spinifexui/frontend/dist/** -whitespace
@@ -19,8 +19,8 @@
  </defs>
  <rect width="900" height="480" fill="url(#bg)" rx="12"/>

-  <text x="24" y="30" font-family="'Segoe UI', system-ui, sans-serif" font-size="11" fill="#a1a1aa" letter-spacing="2.5" font-weight="600">NGINX + ALB · VPC WITH PUBLIC AND PRIVATE SUBNETS</text>
-  <text x="876" y="30" font-family="'SF Mono', ui-monospace, monospace" font-size="10" fill="#71717a" text-anchor="end">10.20.0.0/16 · multi-AZ</text>
+  <text x="24" y="30" font-family="'Segoe UI', system-ui, sans-serif" font-size="11" fill="#a1a1aa" letter-spacing="2.5" font-weight="600">NGINX + ALB · PUBLIC SUBNETS · SPREAD ACROSS NODES</text>
+  <text x="876" y="30" font-family="'SF Mono', ui-monospace, monospace" font-size="10" fill="#71717a" text-anchor="end">10.20.0.0/16 · server_count = 3</text>

  <!-- Internet -->
  <rect x="350" y="54" width="200" height="40" rx="8" fill="url(#inet-grad)" stroke="#3fb950" stroke-width="1.1" stroke-opacity="0.55"/>
@@ -41,58 +41,45 @@
  <line x1="230" y1="184" x2="230" y2="202" stroke="#DC2626" stroke-width="1" stroke-dasharray="3,3" opacity="0.6"/>
  <line x1="670" y1="184" x2="670" y2="202" stroke="#DC2626" stroke-width="1" stroke-dasharray="3,3" opacity="0.6"/>

-  <!-- public_a -->
  <g transform="translate(108, 202)">
-    <rect width="244" height="104" rx="8" fill="url(#pub-grad)" stroke="#3fb950" stroke-width="1.2" stroke-opacity="0.55"/>
+    <rect width="244" height="150" rx="8" fill="url(#pub-grad)" stroke="#3fb950" stroke-width="1.2" stroke-opacity="0.55"/>
    <text x="20" y="24" font-family="'SF Mono', ui-monospace, monospace" font-size="12" fill="#3fb950" font-weight="700">public_a</text>
-    <text x="224" y="24" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#7ee787" text-anchor="end">az-a</text>
+    <text x="224" y="24" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#7ee787" text-anchor="end">10.20.1.0/24</text>
    <line x1="16" y1="32" x2="228" y2="32" stroke="#3fb950" stroke-opacity="0.3" stroke-width="0.5"/>
-
-    <rect x="20" y="42" width="200" height="22" rx="4" fill="#3fb9501a" stroke="#3fb950" stroke-opacity="0.5" stroke-width="0.6"/>
-    <text x="120" y="57" font-family="'SF Mono', ui-monospace, monospace" font-size="10.5" fill="#7ee787" text-anchor="middle" font-weight="600">ALB ENI</text>
-
-    <rect x="20" y="70" width="200" height="22" rx="4" fill="#3fb9501a" stroke="#3fb950" stroke-opacity="0.5" stroke-width="0.6"/>
-    <text x="120" y="85" font-family="'SF Mono', ui-monospace, monospace" font-size="10.5" fill="#7ee787" text-anchor="middle" font-weight="600">NAT GW</text>
+    <rect x="22" y="42" width="200" height="22" rx="4" fill="#3fb9501a" stroke="#3fb950" stroke-opacity="0.5" stroke-width="0.6"/>
+    <text x="122" y="57" font-family="'SF Mono', ui-monospace, monospace" font-size="10.5" fill="#7ee787" text-anchor="middle" font-weight="600">ALB ENI</text>
+    <rect x="22" y="70" width="200" height="36" rx="4" fill="#3fb9501a" stroke="#3fb950" stroke-opacity="0.5" stroke-width="0.6"/>
+    <text x="122" y="86" font-family="'SF Mono', ui-monospace, monospace" font-size="10.5" fill="#7ee787" text-anchor="middle" font-weight="600">nginx-alb-1</text>
+    <text x="122" y="100" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#a1a1aa" text-anchor="middle">public IP · Ubuntu · cloud-init</text>
+    <rect x="22" y="110" width="200" height="36" rx="4" fill="#3fb9501a" stroke="#3fb950" stroke-opacity="0.5" stroke-width="0.6"/>
+    <text x="122" y="126" font-family="'SF Mono', ui-monospace, monospace" font-size="10.5" fill="#7ee787" text-anchor="middle" font-weight="600">nginx-alb-3</text>
+    <text x="122" y="140" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#a1a1aa" text-anchor="middle">public IP · Ubuntu · cloud-init</text>
  </g>

-  <!-- public_b -->
  <g transform="translate(548, 202)">
-    <rect width="244" height="104" rx="8" fill="url(#pub-grad)" stroke="#3fb950" stroke-width="1.2" stroke-opacity="0.55"/>
+    <rect width="244" height="150" rx="8" fill="url(#pub-grad)" stroke="#3fb950" stroke-width="1.2" stroke-opacity="0.55"/>
    <text x="20" y="24" font-family="'SF Mono', ui-monospace, monospace" font-size="12" fill="#3fb950" font-weight="700">public_b</text>
-    <text x="224" y="24" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#7ee787" text-anchor="end">az-b</text>
+    <text x="224" y="24" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#7ee787" text-anchor="end">10.20.2.0/24</text>
    <line x1="16" y1="32" x2="228" y2="32" stroke="#3fb950" stroke-opacity="0.3" stroke-width="0.5"/>
-
-    <rect x="20" y="56" width="200" height="22" rx="4" fill="#3fb9501a" stroke="#3fb950" stroke-opacity="0.5" stroke-width="0.6"/>
-    <text x="120" y="71" font-family="'SF Mono', ui-monospace, monospace" font-size="10.5" fill="#7ee787" text-anchor="middle" font-weight="600">ALB ENI</text>
+    <rect x="22" y="42" width="200" height="22" rx="4" fill="#3fb9501a" stroke="#3fb950" stroke-opacity="0.5" stroke-width="0.6"/>
+    <text x="122" y="57" font-family="'SF Mono', ui-monospace, monospace" font-size="10.5" fill="#7ee787" text-anchor="middle" font-weight="600">ALB ENI</text>
+    <rect x="22" y="70" width="200" height="36" rx="4" fill="#3fb9501a" stroke="#3fb950" stroke-opacity="0.5" stroke-width="0.6"/>
+    <text x="122" y="86" font-family="'SF Mono', ui-monospace, monospace" font-size="10.5" fill="#7ee787" text-anchor="middle" font-weight="600">nginx-alb-2</text>
+    <text x="122" y="100" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#a1a1aa" text-anchor="middle">public IP · Ubuntu · cloud-init</text>
  </g>

-  <!-- SNAT arrow from public_a NAT → private row -->
-  <line x1="230" y1="306" x2="230" y2="340" stroke="#F87171" stroke-width="1.2" stroke-dasharray="3,3" opacity="0.65"/>
-  <polygon points="226,336 234,336 230,344" fill="#F87171" opacity="0.8"/>
-  <rect x="256" y="314" width="240" height="20" rx="10" fill="#0a0a0a" stroke="#F87171" stroke-width="0.6" stroke-opacity="0.55"/>
-  <text x="376" y="328" font-family="'SF Mono', ui-monospace, monospace" font-size="10" fill="#FCA5A5" text-anchor="middle">SNAT for private subnets (egress)</text>
-
-  <!-- private_a -->
-  <g transform="translate(108, 350)">
-    <rect width="244" height="104" rx="8" fill="url(#priv-grad)" stroke="#a855f7" stroke-width="1.2" stroke-opacity="0.55"/>
-    <text x="20" y="24" font-family="'SF Mono', ui-monospace, monospace" font-size="12" fill="#c084fc" font-weight="700">private_a</text>
-    <text x="224" y="24" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#d8b4fe" text-anchor="end">az-a</text>
-    <line x1="16" y1="32" x2="228" y2="32" stroke="#a855f7" stroke-opacity="0.3" stroke-width="0.5"/>
-
-    <rect x="20" y="44" width="200" height="48" rx="5" fill="#a855f71a" stroke="#a855f7" stroke-opacity="0.5" stroke-width="0.6"/>
-    <text x="120" y="64" font-family="'SF Mono', ui-monospace, monospace" font-size="11" fill="#d8b4fe" text-anchor="middle" font-weight="700">nginx_1</text>
-    <text x="120" y="80" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#a1a1aa" text-anchor="middle">Debian 12 · cloud-init</text>
-  </g>
-
-  <!-- private_b -->
-  <g transform="translate(548, 350)">
-    <rect width="244" height="104" rx="8" fill="url(#priv-grad)" stroke="#a855f7" stroke-width="1.2" stroke-opacity="0.55"/>
-    <text x="20" y="24" font-family="'SF Mono', ui-monospace, monospace" font-size="12" fill="#c084fc" font-weight="700">private_b</text>
-    <text x="224" y="24" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#d8b4fe" text-anchor="end">az-b</text>
-    <line x1="16" y1="32" x2="228" y2="32" stroke="#a855f7" stroke-opacity="0.3" stroke-width="0.5"/>
-
-    <rect x="20" y="44" width="200" height="48" rx="5" fill="#a855f71a" stroke="#a855f7" stroke-opacity="0.5" stroke-width="0.6"/>
-    <text x="120" y="64" font-family="'SF Mono', ui-monospace, monospace" font-size="11" fill="#d8b4fe" text-anchor="middle" font-weight="700">nginx_2</text>
-    <text x="120" y="80" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#a1a1aa" text-anchor="middle">Debian 12 · cloud-init</text>
-  </g>
+  <!-- spread placement group -->
+  <rect x="108" y="372" width="684" height="88" rx="8" fill="url(#priv-grad)" stroke="#a855f7" stroke-width="1.2" stroke-opacity="0.55"/>
+  <text x="128" y="394" font-family="'SF Mono', ui-monospace, monospace" font-size="12" fill="#c084fc" font-weight="700">nginx-alb-spread</text>
+  <text x="772" y="394" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#d8b4fe" text-anchor="end">placement group · strategy spread · one server per node</text>
+  <line x1="124" y1="402" x2="776" y2="402" stroke="#a855f7" stroke-opacity="0.3" stroke-width="0.5"/>
+  <rect x="128" y="412" width="204" height="36" rx="5" fill="#a855f71a" stroke="#a855f7" stroke-opacity="0.5" stroke-width="0.6"/>
+  <text x="230" y="428" font-family="'SF Mono', ui-monospace, monospace" font-size="11" fill="#d8b4fe" text-anchor="middle" font-weight="700">node1</text>
+  <text x="230" y="442" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#a1a1aa" text-anchor="middle">one web server</text>
+  <rect x="348" y="412" width="204" height="36" rx="5" fill="#a855f71a" stroke="#a855f7" stroke-opacity="0.5" stroke-width="0.6"/>
+  <text x="450" y="428" font-family="'SF Mono', ui-monospace, monospace" font-size="11" fill="#d8b4fe" text-anchor="middle" font-weight="700">node2</text>
+  <text x="450" y="442" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#a1a1aa" text-anchor="middle">one web server</text>
+  <rect x="568" y="412" width="204" height="36" rx="5" fill="#a855f71a" stroke="#a855f7" stroke-opacity="0.5" stroke-width="0.6"/>
+  <text x="670" y="428" font-family="'SF Mono', ui-monospace, monospace" font-size="11" fill="#d8b4fe" text-anchor="middle" font-weight="700">node3</text>
+  <text x="670" y="442" font-family="'Segoe UI', sans-serif" font-size="9.5" fill="#a1a1aa" text-anchor="middle">one web server</text>
 </svg>
@@ -165,7 +165,7 @@
    "runner": "ci-multi",
    "extra": "",
    "suite": "instancerecovery",
-    "timeout_minutes": 105
+    "timeout_minutes": 125
  },
  {
    "cell": 32,
@@ -190,5 +190,60 @@
    "extra": "",
    "suite": "kvquorum",
    "timeout_minutes": 55
+  },
+  {
+    "cell": 34,
+    "install": "binary",
+    "topo": "single",
+    "net": "dhcp",
+    "host": "oracle-10",
+    "arch": "x86_64",
+    "runner": "ci-single",
+    "extra": "",
+    "suite": ""
+  },
+  {
+    "cell": 35,
+    "install": "binary",
+    "topo": "single",
+    "net": "static",
+    "host": "oracle-10",
+    "arch": "x86_64",
+    "runner": "ci-single",
+    "extra": "reboot-single",
+    "suite": ""
+  },
+  {
+    "cell": 36,
+    "install": "binary",
+    "topo": "single",
+    "net": "nat",
+    "host": "oracle-10",
+    "arch": "x86_64",
+    "runner": "ci-single",
+    "extra": "nat-single",
+    "suite": ""
+  },
+  {
+    "cell": 37,
+    "install": "binary",
+    "topo": "real-multi",
+    "net": "dhcp",
+    "host": "oracle-10",
+    "arch": "x86_64",
+    "runner": "ci-multi",
+    "extra": "",
+    "suite": ""
+  },
+  {
+    "cell": 38,
+    "install": "binary",
+    "topo": "real-multi",
+    "net": "nat",
+    "host": "oracle-10",
+    "arch": "x86_64",
+    "runner": "ci-multi",
+    "extra": "nat-multi",
+    "suite": ""
  }
 ]
@@ -16,6 +16,10 @@
 - Evidence added or rechecked:
 - ADR-0008 lifecycle register / ADR-0001 migration-record / bead updates:

+## Mulga design
+
+<!-- One line, so a reviewer knows whether the design records in the private mulga repository need to change. Write either "Updated: <mulga PR or bead id>" or "Not needed: <reason>". This is a reminder, not a gate. -->
+
 ## Community Contribution License

 - [ ] All community contributions in this pull request are licensed to Mulga under the terms of the [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0). By creating this pull request I represent that I have the right to license these contributions to Mulga under the Apache License 2.0.
@@ -0,0 +1,36 @@
+#!/usr/bin/env python3
+"""Warn when a PR body has no filled "Mulga design" line. Reads PR_BODY; never fails the job."""
+
+import os
+import re
+
+COMMENT_RE = re.compile(r"<!--.*?-->", re.S)
+SECTION_RE = re.compile(r"^#{1,6}\s*Mulga design\s*$", re.I)
+LINE_RE = re.compile(r"^\s*(Updated|Not needed)\s*:\s*(.*?)\s*$", re.I)
+PLACEHOLDER_RE = re.compile(r"<[^<>]*>")
+
+
+def filled(body):
+    inside = False
+    for line in COMMENT_RE.sub("", body or "").splitlines():
+        if SECTION_RE.match(line):
+            inside = True
+        elif inside and line.lstrip().startswith("#"):
+            return False
+        elif inside:
+            m = LINE_RE.match(line)
+            if m and PLACEHOLDER_RE.sub("", m.group(2)).strip():
+                return True
+    return False
+
+
+def main():
+    if filled(os.environ.get("PR_BODY", "")):
+        print("The Mulga design line is filled in.")
+    else:
+        print('::warning title=Mulga design::Add a line under "## Mulga design": "Updated: <mulga PR or bead id>" or "Not needed: <reason>".')
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())
@@ -117,8 +117,9 @@ e2e_suite_timeout() {
    # storage case waits for a pause measured at 4m47s and then holds four
    # minutes to prove nothing moved. Three tests, each taking a node away and
    # bringing it back, and the partition case adds a lease validity, a lease
-    # TTL and a boot on top of everything the other two wait for.
-    instancerecovery) echo "90m" ;;
+    # TTL and a boot on top of everything the other two wait for. The data
+    # test lets its guest age past the pending watchdog twice, about 20 more.
+    instancerecovery) echo "110m" ;;
    # lbrecovery pays for a load balancer twice: once to stand one up with two
    # backends placed around it, and once for the recovery itself — a stale
    # owner, a lease, a boot, HAProxy's own start, and its health checks against
@@ -0,0 +1,23 @@
+import unittest
+
+from mulga_design_check import filled
+
+
+class FilledTest(unittest.TestCase):
+    def test_accepts_either_form(self):
+        self.assertTrue(filled("## Mulga design\nUpdated: mulga#275"))
+        self.assertTrue(filled("## Mulga design\n\nnot needed: internal refactor\n## Next"))
+
+    def test_rejects_unanswered(self):
+        self.assertFalse(filled(""))
+        self.assertFalse(filled("## Mulga design\n<!-- Updated: x -->"))
+        self.assertFalse(filled("## Mulga design\nUpdated: <mulga PR or bead id>"))
+        self.assertFalse(filled("## Mulga design\nUpdated:"))
+
+    def test_scoped_to_the_section(self):
+        self.assertFalse(filled("Updated: mulga#1"))
+        self.assertFalse(filled("## Mulga design\n## Other\nUpdated: mulga#1"))
+
+
+if __name__ == "__main__":
+    unittest.main()
@@ -18,6 +18,11 @@ workflows:
        - 'actions/create-github-app-token@v3.3.0'
        - 'actions/upload-artifact@v7.0.2'
        - 'hashicorp/setup-terraform@v4.0.1'
+    '.github/workflows/e2e-cloudvendor-vultr.yml':
+        - 'actions/checkout@v7.0.1'
+        - 'actions/create-github-app-token@v3.3.0'
+        - 'actions/upload-artifact@v7.0.2'
+        - 'hashicorp/setup-terraform@v4.0.1'
    '.github/workflows/e2e-ddil.yml':
        - 'actions/checkout@v7.0.1'
        - 'actions/create-github-app-token@v3.3.0'
@@ -45,6 +50,8 @@ workflows:
    '.github/workflows/image-scripts.yml':
        - 'actions/checkout@v7.0.1'
        - 'actions/setup-go@v7.0.0'
+    '.github/workflows/mulga-design.yml':
+        - 'actions/checkout@v7.0.1'
    '.github/workflows/perf-3node.yml':
        - 'actions/checkout@v7.0.1'
        - 'actions/create-github-app-token@v3.3.0'
@@ -73,6 +80,7 @@ workflows:
    '.github/workflows/ui.yml':
        - 'actions/checkout@v7.0.1'
        - 'actions/setup-node@v7.1.0'
+        - 'actions/upload-artifact@v7.0.2'
        - 'pnpm/action-setup@v6.1.0'
 dependencies:
    'actions/cache@v6.1.0':
@@ -2,7 +2,8 @@ name: Apt Packages

 # Path-filtered like image-scripts.yml: this resolves setup.sh's runtime package
 # list against each supported suite's real indices, so a renamed, dropped, or
-# ambiguously virtual package fails here rather than on a user's machine.
+# ambiguously virtual package fails here rather than on a user's machine. The
+# dnf job does the same for Oracle Linux, through setup.sh's own repo setup.
 on:
  push:
    branches:
@@ -11,6 +12,7 @@ on:
    paths:
      - scripts/setup.sh
      - scripts/lint-apt-packages.sh
+      - scripts/lint-dnf-packages.sh
      - .github/workflows/apt-packages.yml
  pull_request:
    branches:
@@ -19,6 +21,7 @@ on:
    paths:
      - scripts/setup.sh
      - scripts/lint-apt-packages.sh
+      - scripts/lint-dnf-packages.sh
      - .github/workflows/apt-packages.yml
  workflow_dispatch:

@@ -44,3 +47,14 @@ jobs:

      - name: Resolve runtime package list
        run: ./scripts/lint-apt-packages.sh ${{ matrix.image }}
+
+  dnf_packages:
+    name: Resolve dnf list (oraclelinux:10)
+    runs-on: ubuntu-26.04
+
+    steps:
+      - name: Checkout Code
+        uses: actions/checkout@v7.0.1
+
+      - name: Resolve runtime package list
+        run: ./scripts/lint-dnf-packages.sh oraclelinux:10
@@ -102,9 +102,9 @@ jobs:
        env:
          GH_TOKEN: ${{ github.token }}

-      - name: Build microVM artifacts (lb-agent + kernel + initramfs)
+      - name: Build microVM artifacts (lb-agent + kernel + initramfs + qboot + SQS store and worker images)
        if: steps.version.outputs.exists != 'true'
-        run: cd spinifex && make build-microvm-image
+        run: cd spinifex && make build-microvm-image build-sqs-store-microvm-image build-sqs-worker-microvm-image

      - name: Build amd64 Tarball
        if: steps.version.outputs.exists != 'true'
@@ -144,6 +144,7 @@ jobs:
              --notes-file "$RUNNER_TEMP/notes.md" \
              "dist/spinifex-$TAG-linux-amd64.tar.gz" \
              "dist/spinifex-$TAG-linux-amd64.tar.gz.sha256" \
+              dist/qboot-*.tar.gz \
              "scripts/setup.sh"

      # The assertion that matters. A prerelease must never become the latest
@@ -0,0 +1,142 @@
+name: E2E Cloud Vendor Vultr
+
+# On demand only. Deploys a ref to Vultr with deploy-vultr.sh, runs the same
+# published Terraform workbooks as the hypervisor nightly and OCI, and destroys
+# the deployment, failing if anything is left billing.
+on:
+  workflow_dispatch:
+    inputs:
+      nodes:
+        description: "1 or 3 nodes"
+        required: false
+        default: "1"
+        type: choice
+        options: ["1", "3"]
+      plan:
+        description: "Vultr plan: vx1-g-16c-64g (VM) or vbm-8c-64gb-amd (bare metal)"
+        required: false
+        default: vx1-g-16c-64g
+      region:
+        description: "Vultr region"
+        required: false
+        default: ord
+      workbooks:
+        description: "Space-separated workbooks (blank = the driver's full list)"
+        required: false
+      source:
+        description: "tree (build this ref) or release (the published installer)"
+        required: false
+        default: tree
+        type: choice
+        options: [tree, release]
+      channel:
+        description: "With source=release: latest, or dev for the newest prerelease"
+        required: false
+        default: latest
+      keep_on_fail:
+        description: "Leave a failed deployment up to inspect. It keeps billing."
+        required: false
+        type: boolean
+        default: false
+      runner:
+        description: "Runner label (blank = the VULTR_E2E_RUNNER variable, else ubuntu-26.04)"
+        required: false
+
+permissions:
+  contents: read
+
+# Deployments are named per run, but the account's reserved-IP quota is shared,
+# so a second dispatch queues rather than competing with the first.
+concurrency:
+  group: e2e-vultr
+  cancel-in-progress: false
+
+jobs:
+  vultr:
+    # Hosted by default: the job only drives Terraform and SSH. The API key's
+    # access control must admit this runner, so a hosted runner needs a key
+    # whose list holds 0.0.0.0/0; a self-hosted runner can use a fixed address.
+    runs-on: ${{ inputs.runner || vars.VULTR_E2E_RUNNER || 'ubuntu-26.04' }}
+    timeout-minutes: 240
+    env:
+      VULTR_NAME: e2e-${{ github.run_id }}
+      VULTR_NODES: ${{ inputs.nodes }}
+      VULTR_PLAN: ${{ inputs.plan }}
+      VULTR_REGION: ${{ inputs.region }}
+      VULTR_SOURCE: ${{ inputs.source }}
+      VULTR_CHANNEL: ${{ inputs.channel }}
+      VULTR_ARTIFACT_DIR: ${{ github.workspace }}/e2e-vultr
+      SPX_REF: ${{ github.ref_name }}
+    steps:
+      - uses: actions/checkout@v7.0.1
+        with:
+          path: spinifex
+          fetch-depth: 0
+
+      - uses: actions/create-github-app-token@v3.3.0
+        id: app-token
+        with:
+          client-id: ${{ secrets.APP_ID }}
+          private-key: ${{ secrets.APP_PRIVATE_KEY }}
+          repositories: mulga
+
+      # build-install-artifacts.sh lives in mulga; the workbooks and their driver
+      # are in spinifex, so only the one directory is needed here.
+      - uses: actions/checkout@v7.0.1
+        with:
+          repository: mulgadc/mulga
+          path: mulga
+          token: ${{ steps.app-token.outputs.token }}
+          sparse-checkout: scripts/tofu-cluster
+
+      - uses: hashicorp/setup-terraform@v4.0.1
+        with:
+          terraform_wrapper: false
+
+      - name: Run Vultr
+        working-directory: spinifex/scripts/terraform/vultr-spx
+        env:
+          VULTR_API_KEY: ${{ secrets.VULTR_API_KEY }}
+          MULGA_ROOT: ${{ github.workspace }}/mulga
+          # Through the environment, never interpolated into the script: a
+          # dispatch input is untrusted text to the shell.
+          WORKBOOKS_INPUT: ${{ inputs.workbooks }}
+          VULTR_KEEP_ON_FAIL: ${{ inputs.keep_on_fail && '1' || '0' }}
+        run: |
+          set -uo pipefail
+          if [ -n "$WORKBOOKS_INPUT" ]; then export WORKBOOKS="$WORKBOOKS_INPUT"; fi
+          ./e2e-vultr.sh 2>&1 | tee "$RUNNER_TEMP/e2e-vultr.log"
+          exit "${PIPESTATUS[0]}"
+
+      # always(): a cancelled or timed-out job is the one that would leave
+      # servers and reserved IPs billing. destroy works from this run's state.
+      - name: Sweep anything left behind
+        if: always() && !inputs.keep_on_fail
+        working-directory: spinifex/scripts/terraform/vultr-spx
+        env:
+          VULTR_API_KEY: ${{ secrets.VULTR_API_KEY }}
+        run: ./e2e-vultr.sh --sweep-only
+
+      - name: Publish the verdict
+        if: always()
+        run: |
+          {
+            echo "### Vultr: ${VULTR_NODES} x ${VULTR_PLAN} in ${VULTR_REGION}, ${VULTR_SOURCE} build of ${SPX_REF}"
+            echo '```'
+            cat "$VULTR_ARTIFACT_DIR/verdict.txt" 2>/dev/null || echo "no verdict written"
+            echo '```'
+          } >> "$GITHUB_STEP_SUMMARY"
+
+      # Logs and the verdict only. The throwaway SSH key and the Terraform state
+      # under ~/.spinifex never reach an artifact.
+      - name: Upload the run
+        if: always()
+        uses: actions/upload-artifact@v7.0.2
+        with:
+          name: e2e-vultr-${{ github.run_id }}
+          path: |
+            ${{ runner.temp }}/e2e-vultr.log
+            ${{ github.workspace }}/e2e-vultr/*.log
+            ${{ github.workspace }}/e2e-vultr/verdict.txt
+            ${{ github.workspace }}/e2e-vultr/node*/*.log
+          retention-days: 14
@@ -10,7 +10,7 @@ on:
  workflow_dispatch:
    inputs:
      cells:
-        description: "Comma-separated cell numbers to run, no spaces (blank = all). 1-3, 8-10, 17-19 tofu permutation matrix, 20 baremetal, 21-22 ISO, 23 predastore throughput, 24 predastore fault injection, 25-27 service suites (eks, ecs, rds), 28 guest disk performance, 29 storage fault injection, 30 routed NAT on three nodes, 31 instance recovery on three nodes, 32 load balancer recovery on three nodes, 33 KV bucket quorum under node loss on three nodes"
+        description: "Comma-separated cell numbers to run, no spaces (blank = all). 1-3, 8-10, 17-19 tofu permutation matrix, 34-36 Oracle Linux single node (basic, reboot, routed NAT), 37-38 Oracle Linux on three nodes (multi-node subset with IPsec, routed NAT), 20 baremetal, 21-22 ISO, 23 predastore throughput, 24 predastore fault injection, 25-27 service suites (eks, ecs, rds), 28 guest disk performance, 29 storage fault injection, 30 routed NAT on three nodes, 31 instance recovery on three nodes, 32 load balancer recovery on three nodes, 33 KV bucket quorum under node loss on three nodes"
        required: false

 permissions:
@@ -33,6 +33,7 @@ concurrency:
 env:
  GIT_BRANCH: ${{ github.ref_name }}
  UBUNTU_GOLD_IMAGE: /mnt/disk1/gold-images/spinifex-image-builder-ubuntu-26.04.qcow2
+  ORACLE_GOLD_IMAGE: /mnt/disk1/gold-images/spinifex-image-builder-oracle-10.qcow2
  SPINIFEX_OTLP_ENDPOINT: ${{ vars.SPINIFEX_OTLP_ENDPOINT || 'http://192.168.1.13:8200' }}
  SPINIFEX_OTEL_ATTRS: ci.run_id=${{ github.run_id }},ci.workflow=${{ github.workflow }}

@@ -195,7 +196,8 @@ jobs:
          # back onto the debian gold.
          TOFU_VAR_ARGS=""
          case "$MATRIX_HOST" in
-            ubuntu-*) TOFU_VAR_ARGS="-var=host_image_override=${UBUNTU_GOLD_IMAGE}" ;;
+            ubuntu-*) TOFU_VAR_ARGS="-var=host_image_override=${UBUNTU_GOLD_IMAGE} -var=host_os=ubuntu" ;;
+            oracle-*) TOFU_VAR_ARGS="-var=host_image_override=${ORACLE_GOLD_IMAGE} -var=host_os=oracle" ;;
          esac

          source "${GITHUB_WORKSPACE}/spinifex/.github/scripts/ci-fmt.sh"
@@ -687,7 +689,7 @@ jobs:
              'echo "host: $(hostname)"; echo "loadavg: $(cat /proc/loadavg)";
               grep -E "^(MemTotal|MemAvailable|SwapFree):" /proc/meminfo;
               echo "cpu(steal is field 8): $(grep "^cpu " /proc/stat)";
-               echo "running qemu: $(pgrep -c qemu-system-x86 || echo 0)"' 2>&1
+               echo "running qemu: $(pgrep -c "qemu-system-x86|qemu-kvm" || echo 0)"' 2>&1
          } >> "${ARTIFACT_DIR}/node-load.txt"
          exit 0

@@ -858,7 +860,8 @@ jobs:
          # different image path than apply did.
          TOFU_VAR_ARGS=""
          case "$MATRIX_HOST" in
-            ubuntu-*) TOFU_VAR_ARGS="-var=host_image_override=${UBUNTU_GOLD_IMAGE}" ;;
+            ubuntu-*) TOFU_VAR_ARGS="-var=host_image_override=${UBUNTU_GOLD_IMAGE} -var=host_os=ubuntu" ;;
+            oracle-*) TOFU_VAR_ARGS="-var=host_image_override=${ORACLE_GOLD_IMAGE} -var=host_os=oracle" ;;
          esac
          source "${GITHUB_WORKSPACE}/spinifex/.github/scripts/ci-fmt.sh"
          tofu workspace select -or-create "${SPINIFEX_CI_ENV}" >> "$TOFU_LOG" 2>&1
@@ -536,6 +536,13 @@ jobs:
          log-dir: ${{ env.ARTIFACT_DIR }}
          title: "E2E failure analysis"

+      # The WAN capture is most of the artifact and only matters when a run
+      # failed. Anything short of a clean pass (suite rc, a failed setup or
+      # collect step, a missing rc) keeps it.
+      - name: Drop the packet capture from a passing single-node run
+        if: always() && steps.e2e.outputs.rc == '0' && job.status == 'success'
+        run: rm -f "${ARTIFACT_DIR}"/net-diag/wan.pcap
+
      - uses: actions/upload-artifact@v7.0.2
        if: always()
        with:
@@ -990,6 +997,13 @@ jobs:
          log-dir: ${{ env.ARTIFACT_DIR }}
          title: "Multi-Node E2E failure analysis"

+      # The WAN captures are 90% of the artifact and only matter when a run
+      # failed. Anything short of a clean pass (suite rc, a failed setup or
+      # collect step, a missing rc) keeps them.
+      - name: Drop the packet captures from a passing multi-node run
+        if: always() && steps.e2e.outputs.rc == '0' && job.status == 'success'
+        run: rm -f "${ARTIFACT_DIR}"/net-diag-node*/wan.pcap
+
      - uses: actions/upload-artifact@v7.0.2
        if: always()
        with:
@@ -0,0 +1,33 @@
+name: Mulga design
+
+# Reminds the author to say whether the Mulga design records need to change.
+# It reads only the pull request body, uses no secrets, and warns without ever
+# failing: the check is a prompt, not a gate.
+
+on:
+  pull_request:
+    branches:
+      - dev
+    types: [opened, edited, reopened]
+
+permissions:
+  contents: read
+
+jobs:
+  mulga_design:
+    name: Mulga design line
+    if: ${{ !endsWith(github.actor, '[bot]') }}
+    runs-on: ubuntu-26.04
+    steps:
+      - name: Checkout Code
+        uses: actions/checkout@v7.0.1
+        with:
+          sparse-checkout: .github/scripts
+
+      - name: Check the Mulga design line
+        env:
+          PR_BODY: ${{ github.event.pull_request.body }}
+        working-directory: .github/scripts
+        run: |
+          python3 -m unittest -q test_mulga_design_check
+          python3 mulga_design_check.py
@@ -57,8 +57,8 @@ jobs:
          go-version-file: spinifex/go.mod
          cache: false

-      - name: Build microVM artifacts (lb-agent + kernel + initramfs)
-        run: cd spinifex && make build-microvm-image
+      - name: Build microVM artifacts (lb-agent + kernel + initramfs + qboot + SQS store and worker images)
+        run: cd spinifex && make build-microvm-image build-sqs-store-microvm-image build-sqs-worker-microvm-image

      - name: Build amd64 Tarball
        env:
@@ -72,6 +72,7 @@ jobs:
          path: |
            spinifex/dist/spinifex-${{ github.ref_name }}-linux-amd64.tar.gz
            spinifex/dist/spinifex-${{ github.ref_name }}-linux-amd64.tar.gz.sha256
+            spinifex/dist/qboot-*.tar.gz
            spinifex/scripts/setup.sh

  distro-arm64:
@@ -119,9 +119,6 @@ jobs:
      - name: Build Spinifex
        run: make -C spinifex build

-      - name: Manifest check + drift lint
-        run: make -C spinifex manifest-check manifest-lint
-
      - name: Run Tests with Coverage
        env:
          SPINIFEX_REQUIRE_CONFORMANCE_TOOLS: "1"
@@ -7,6 +7,11 @@ on:
      - dev
    paths:
      - spinifex/runtime/roles/spinifexui/frontend/**
+      - spinifex/runtime/roles/spinifexui/*.go
+      - spinifex/operator/admin/instance_metrics*.go
+      - spinifex/operator/admin/live_streams*.go
+      - spinifex/runtime/roles/qmpcollector/**
+      - contracts/telemetry/**
      - .github/workflows/ui.yml
  pull_request:
    branches:
@@ -14,6 +19,11 @@ on:
      - dev
    paths:
      - spinifex/runtime/roles/spinifexui/frontend/**
+      - spinifex/runtime/roles/spinifexui/*.go
+      - spinifex/operator/admin/instance_metrics*.go
+      - spinifex/operator/admin/live_streams*.go
+      - spinifex/runtime/roles/qmpcollector/**
+      - contracts/telemetry/**
      - .github/workflows/ui.yml
  workflow_dispatch:

@@ -69,3 +79,85 @@ jobs:
            echo "$drift"
            exit 1
          fi
+
+  # The pinned image lives in one file that scripts/ui-playwright.sh also reads,
+  # so make ui-baselines and the job below draw screenshots in the same image.
+  ui_e2e_image:
+    name: UI End-to-End Image
+    runs-on: ubuntu-26.04
+    outputs:
+      image: ${{ steps.read.outputs.image }}
+
+    steps:
+      - name: Checkout Code
+        uses: actions/checkout@v7.0.1
+        with:
+          sparse-checkout: spinifex/runtime/roles/spinifexui/frontend/playwright.image
+          sparse-checkout-cone-mode: false
+
+      - name: Read Playwright Image
+        id: read
+        run: echo "image=$(tr -d '[:space:]' < spinifex/runtime/roles/spinifexui/frontend/playwright.image)" >> "$GITHUB_OUTPUT"
+
+  # The browser suite: the console against a faked API on Vite dev, plus the
+  # built dist under the production CSP. It runs in Playwright's own image, which
+  # carries the browsers, so screenshots rasterise as they did for the baselines.
+  ui_e2e:
+    name: UI End-to-End
+    needs: ui_e2e_image
+    runs-on: ubuntu-26.04
+    timeout-minutes: 20
+    container:
+      image: ${{ needs.ui_e2e_image.outputs.image }}
+      options: --ipc=host
+    defaults:
+      run:
+        working-directory: spinifex/runtime/roles/spinifexui/frontend
+
+    steps:
+      - name: Checkout Code
+        uses: actions/checkout@v7.0.1
+        with:
+          sparse-checkout: spinifex/runtime/roles/spinifexui/frontend
+
+      - name: Setup pnpm
+        uses: pnpm/action-setup@v6.1.0
+        with:
+          package_json_file: spinifex/runtime/roles/spinifexui/frontend/package.json
+
+      - name: Setup Node.js
+        uses: actions/setup-node@v7.1.0
+        with:
+          cache: "pnpm"
+          cache-dependency-path: spinifex/runtime/roles/spinifexui/frontend/pnpm-lock.yaml
+          node-version-file: spinifex/runtime/roles/spinifexui/frontend/.nvmrc
+
+      - name: Install Dependencies
+        run: pnpm install --frozen-lockfile
+
+      # The image's browsers belong to one Playwright release, so a lockfile
+      # bump without an image bump would otherwise fail on a missing executable.
+      - name: Check Playwright Matches the Image
+        env:
+          IMAGE: ${{ needs.ui_e2e_image.outputs.image }}
+        run: |
+          version=$(pnpm exec playwright --version | awk '{print $2}')
+          case "$IMAGE" in
+            *":v$version-"*) ;;
+            *) echo "::error::Lockfile has Playwright $version but the image is $IMAGE. Update playwright.image."; exit 1 ;;
+          esac
+
+      - name: Build Frontend
+        run: pnpm build
+
+      - name: Run End-to-End Suite
+        run: pnpm e2e
+
+      - name: Upload Screenshot Diffs
+        if: failure()
+        uses: actions/upload-artifact@v7.0.2
+        with:
+          name: ui-e2e-test-results
+          path: spinifex/runtime/roles/spinifexui/frontend/test-results/
+          if-no-files-found: ignore
+          retention-days: 14
@@ -19,6 +19,15 @@ bin/
 dist/
 build/microvm/vmlinuz
 build/microvm/initramfs.cpio.gz
+build/microvm/qboot.rom
+build/microvm/qboot.LICENSE
+build/microvm/sqs-store/vmlinuz
+build/microvm/sqs-store/initramfs.cpio.gz
+build/microvm/sqs-store/.stage.*/
+build/microvm/sqs-worker/vmlinuz
+build/microvm/sqs-worker/initramfs.cpio.gz
+build/microvm/sqs-worker/.stage.*/
+build/qboot/
 !spinifex/runtime/roles/spinifexui/frontend/dist/
 tmp/
 *.test
@@ -56,6 +56,19 @@ build-ui:
 	@echo -e "\n....Building spinifex-ui frontend...."
 	cd spinifex/runtime/roles/spinifexui/frontend && pnpm build

+# Playwright suite for the console, against Vite dev and the built dist under
+# the console's CSP. Not part of preflight; needs `pnpm exec playwright install chromium`.
+test-ui-e2e: build-ui
+	cd spinifex/runtime/roles/spinifexui/frontend && pnpm e2e
+
+# The same suite inside the pinned Playwright image CI uses (podman or docker),
+# so screenshots compare as they will on the runner. ui-baselines redraws them.
+ui-e2e-container:
+	scripts/ui-playwright.sh compare
+
+ui-baselines:
+	scripts/ui-playwright.sh baselines
+
 # GO commands
 VERSION ?= $(shell git describe --tags --always --dirty --exclude '*-dev')
 COMMIT  ?= $(shell git rev-parse --short HEAD)
@@ -140,22 +153,76 @@ $(MICROVM_ARTIFACTS) &: $(MICROVM_INPUTS)
 bin/lb-agent:
 	$(MAKE) build-lb-agent

-build-microvm-image: $(MICROVM_ARTIFACTS) ## Build microVM kernel + initramfs (incremental — skips when up to date)
+# The SQS store microVM image: a separate artifact from the LB image above, not
+# installed or released by any target yet.
+SQS_STORE_MICROVM_OUT_DIR := $(MICROVM_OUT_DIR)/sqs-store
+SQS_STORE_MICROVM_ARTIFACTS := $(SQS_STORE_MICROVM_OUT_DIR)/vmlinuz $(SQS_STORE_MICROVM_OUT_DIR)/initramfs.cpio.gz
+SQS_STORE_MICROVM_INPUTS := scripts/build-sqs-store-microvm-image.sh scripts/lib/sqs-store-microvm-publish.sh $(SQS_STORE_MICROVM_OUT_DIR)/init.sh $(SQS_STORE_MICROVM_OUT_DIR)/inittab $(addprefix $(SQS_STORE_MICROVM_OUT_DIR)/,sqs-store-lib.sh sqs-store-prepare sqs-store-run sqs-store-stop) bin/nats-server

-install-microvm: $(MICROVM_ARTIFACTS) ## Install microVM artifacts to /usr/share/spinifex/microvm/
+$(SQS_STORE_MICROVM_ARTIFACTS) &: $(SQS_STORE_MICROVM_INPUTS)
+	./scripts/build-sqs-store-microvm-image.sh
+
+# The dedicated SQS store server, built from the go.mod-pinned module so the
+# image ships exactly the reviewed version. Static for the musl guest.
+bin/nats-server: go.mod go.sum
+	@mkdir -p bin
+	GOWORK=off CGO_ENABLED=0 GOOS=linux GOARCH=amd64 go build -trimpath -ldflags '-s -w' -o $@ github.com/nats-io/nats-server/v2
+
+build-sqs-store-microvm-image: $(SQS_STORE_MICROVM_ARTIFACTS) ## Build the SQS store microVM kernel and initramfs
+
+# The shared SQS worker microVM image: the store image's /init and kernel with
+# the worker binary instead of nats-server. Not installed or released yet.
+SQS_WORKER_MICROVM_OUT_DIR := $(MICROVM_OUT_DIR)/sqs-worker
+SQS_WORKER_MICROVM_ARTIFACTS := $(SQS_WORKER_MICROVM_OUT_DIR)/vmlinuz $(SQS_WORKER_MICROVM_OUT_DIR)/initramfs.cpio.gz
+SQS_WORKER_MICROVM_INPUTS := scripts/build-sqs-store-microvm-image.sh scripts/lib/sqs-store-microvm-publish.sh $(SQS_STORE_MICROVM_OUT_DIR)/init.sh $(SQS_WORKER_MICROVM_OUT_DIR)/inittab $(addprefix $(SQS_WORKER_MICROVM_OUT_DIR)/,sqs-worker-lib.sh sqs-worker-prepare sqs-worker-ready sqs-worker-stop) bin/sqs-worker
+
+$(SQS_WORKER_MICROVM_ARTIFACTS) &: $(SQS_WORKER_MICROVM_INPUTS)
+	SQS_MICROVM_PROFILE=worker ./scripts/build-sqs-store-microvm-image.sh
+
+bin/sqs-worker: go.mod go.sum $(shell find cmd/sqs-worker spinifex/agents/sqs -name '*.go' ! -name '*_test.go')
+	@mkdir -p bin
+	GOWORK=off CGO_ENABLED=0 GOOS=linux GOARCH=amd64 go build -trimpath -ldflags '-s -w' -o $@ ./cmd/sqs-worker
+
+build-sqs-worker-microvm-image: $(SQS_WORKER_MICROVM_ARTIFACTS) ## Build the shared SQS worker microVM kernel and initramfs
+
+# qboot is the microvm firmware. The source tarball ships with each release
+# because qboot is GPL-2.0.
+QBOOT_COMMIT := $(shell sed -n 's/^QBOOT_COMMIT="\(.*\)"$$/\1/p' scripts/build-qboot.sh)
+QBOOT_SOURCE := build/qboot/qboot-$(QBOOT_COMMIT).tar.gz
+QBOOT_ARTIFACTS := $(MICROVM_OUT_DIR)/qboot.rom $(MICROVM_OUT_DIR)/qboot.LICENSE $(QBOOT_SOURCE)
+
+$(QBOOT_ARTIFACTS) &: scripts/build-qboot.sh
+	./scripts/build-qboot.sh
+
+build-microvm-image: $(MICROVM_ARTIFACTS) $(QBOOT_ARTIFACTS) ## Build microVM kernel, initramfs and qboot firmware (incremental — skips when up to date)
+
+install-sqs-store-microvm: $(SQS_STORE_MICROVM_ARTIFACTS) ## Install the SQS store microVM image to /usr/share/spinifex/microvm/sqs-store/
+	sudo install -d /usr/share/spinifex/microvm/sqs-store
+	sudo rm -f /usr/share/spinifex/microvm/sqs-store/initramfs.cpio.gz
+	sudo install -m 0644 $(SQS_STORE_MICROVM_OUT_DIR)/vmlinuz /usr/share/spinifex/microvm/sqs-store/vmlinuz
+	sudo install -m 0644 $(SQS_STORE_MICROVM_OUT_DIR)/initramfs.cpio.gz /usr/share/spinifex/microvm/sqs-store/initramfs.cpio.gz
+
+install-sqs-worker-microvm: $(SQS_WORKER_MICROVM_ARTIFACTS) ## Install the SQS worker microVM image to /usr/share/spinifex/microvm/sqs-worker/
+	sudo install -d /usr/share/spinifex/microvm/sqs-worker
+	sudo rm -f /usr/share/spinifex/microvm/sqs-worker/initramfs.cpio.gz
+	sudo install -m 0644 $(SQS_WORKER_MICROVM_OUT_DIR)/vmlinuz /usr/share/spinifex/microvm/sqs-worker/vmlinuz
+	sudo install -m 0644 $(SQS_WORKER_MICROVM_OUT_DIR)/initramfs.cpio.gz /usr/share/spinifex/microvm/sqs-worker/initramfs.cpio.gz
+
+install-microvm: $(MICROVM_ARTIFACTS) $(QBOOT_ARTIFACTS) ## Install microVM artifacts to /usr/share/spinifex/microvm/
 	sudo install -d /usr/share/spinifex/microvm
 	sudo install -m 0644 $(MICROVM_OUT_DIR)/vmlinuz /usr/share/spinifex/microvm/vmlinuz
 	sudo install -m 0644 $(MICROVM_OUT_DIR)/initramfs.cpio.gz /usr/share/spinifex/microvm/initramfs.cpio.gz
+	sudo install -m 0644 $(MICROVM_OUT_DIR)/qboot.rom /usr/share/spinifex/microvm/qboot.rom
+	sudo install -m 0644 $(MICROVM_OUT_DIR)/qboot.LICENSE /usr/share/spinifex/microvm/qboot.LICENSE

-# the pre-commit gate: manifest checks, lint, vuln, and the unit and e2e-harness tiers.
+# the pre-commit gate: lint, vuln, and the unit and e2e-harness tiers.
 # integration and race tests skipped to keep quick, they run in CI.
 #
 # -j2 overlaps the only two heavy gates, lint and test-cover: alone each leaves
 # most of the machine idle, so pairing them cuts a cold run by about a quarter.
 # -Otarget keeps each gate's output in one block instead of interleaving them.
-# manifest-check manifest-lint test-package-check removed, can add back later
 preflight:
-	@$(MAKE) --no-print-directory -j2 -Otarget QUIET=1 lint govulncheck test-cover diff-coverage test-harness test-build-scripts manifest-check manifest-lint
+	@$(MAKE) --no-print-directory -j2 -Otarget QUIET=1 lint govulncheck test-cover diff-coverage test-harness test-build-scripts
 	@echo -e "\n ✅ Preflight passed — safe to commit."

 # Shell suites + shellcheck for build/scripts/, the systemd-unit helpers that
@@ -167,7 +234,7 @@ preflight:
 # reason: between them they install the software, wire OVN, erase the state and
 # remove the install, so a shellcheck warning in one is an outage or a wrong
 # deletion, not a style note.
-NODE_SCRIPTS := scripts/setup.sh scripts/setup-ovn.sh scripts/install-node.sh scripts/smoke-test.sh scripts/node-reset.sh scripts/uninstall-spx.sh
+NODE_SCRIPTS := scripts/setup.sh scripts/setup-ovn.sh scripts/install-node.sh scripts/smoke-test.sh scripts/node-reset.sh scripts/uninstall-spx.sh scripts/lib/distro.sh scripts/lib/stage-sqs-store-microvm.sh

 test-build-scripts:
 	@echo -e "\n....Running build/scripts/**/*_test.sh...."
@@ -175,13 +242,14 @@ test-build-scripts:
 		echo "-- $$t"; \
 		sh "$$t" || exit 1; \
 	done
-	@echo -e "\n....Running scripts/*_test.sh...."
-	@for t in $$(find scripts -maxdepth 1 -name '*_test.sh' | sort); do \
+	@echo -e "\n....Running scripts/*_test.sh, scripts/lib/*_test.sh and scripts/appliance/*_test.sh...."
+	@for t in $$( (find scripts -maxdepth 1 -name '*_test.sh'; find scripts/lib scripts/appliance -name '*_test.sh') | sort); do \
 		echo "-- $$t"; \
 		sh "$$t" || exit 1; \
 	done
 	@echo -e "\n....Running shellcheck over build/scripts/**/*.sh + the node scripts...."
 	shellcheck -S warning $$(find build/scripts -name '*.sh' | sort) $(NODE_SCRIPTS)
+	shellcheck -S warning -x scripts/build-oracle-appliance.sh $$(find scripts/appliance -name '*.sh' | sort)
 	@echo "  test-build-scripts ok"

 # E2E harness unit tests. Build-tagged `e2e` so they're skipped by the
@@ -228,23 +296,6 @@ test-segscan-oracle:
 	@echo -e "\n....Running segscan storage oracle test...."
 	$(_Q)LOG_IGNORE=1 go test -tags=integration,segscanoracle -timeout 120s ./tests/integration/... -run TestSegscanOracle $(_RACEQ)

-# Validate docs/service-interfaces.yaml. Schema check + cross-reference
-# of services/suites/fixtures + on-disk path existence.
-manifest-check:
-	@echo -e "\n....Checking service-interfaces.yaml...."
-	@go run ./tests/e2e/manifest-check/cmd/manifest-check -repo-root . -manifest docs/service-interfaces.yaml
-
-# Drift guards: direct-create fixture lint + NATS subject lint,
-# ratcheted against tests/e2e/manifest-lint/baseline.txt. Fails only on NEW
-# drift beyond the baseline.
-manifest-lint:
-	@echo -e "\n....Linting manifest drift (fixtures + subjects)...."
-	@go run ./tests/e2e/manifest-lint/cmd/manifest-lint -repo-root .
-
-# Accept current drift into the baseline. Run after an intentional change.
-manifest-lint-update:
-	@go run ./tests/e2e/manifest-lint/cmd/manifest-lint -repo-root . -update
-
 # Run unit tests
 test:
 	@echo -e "\n....Running tests for $(GO_PROJECT_NAME)...."
@@ -333,14 +384,20 @@ fuzz:
 # Microvm artifacts are reinstalled when they already exist on disk — the rule's
 # input timestamps drive a rebuild only if anything actually changed. On a fresh
 # checkout (no build/microvm/vmlinuz yet) the install step is skipped; run
-# `make install-microvm` explicitly the first time.
+# `make install-microvm` explicitly the first time. They go in before the
+# preflight, which checks the microvm firmware.
 deploy: build
 	sudo install -m 755 bin/spx /usr/local/bin/spx
+	@if [ -f $(MICROVM_OUT_DIR)/vmlinuz ]; then \
+		$(MAKE) install-microvm; \
+	else \
+		echo "[deploy] microvm artifacts absent — run 'make install-microvm' for first-time setup"; \
+	fi
 	@if [ "$${SKIP_ASSET_PREFLIGHT:-}" != "1" ]; then \
 		if ! sudo /usr/local/bin/spx admin preflight; then \
 			echo ""; \
-			echo "[deploy] host-asset preflight failed: this node's helper scripts or sudoers"; \
-			echo "  grants don't match the binary just installed. Remediate with:"; \
+			echo "[deploy] host-asset preflight failed: this node's helper scripts, sudoers"; \
+			echo "  grants, QEMU or microvm firmware don't match the binary just installed. Remediate with:"; \
 			echo "    scripts/update-nodes.sh   (from the mulga root)"; \
 			echo "  or:"; \
 			echo "    make reinstall"; \
@@ -350,11 +407,6 @@ deploy: build
 	else \
 		echo "[deploy] SKIP_ASSET_PREFLIGHT=1 — skipping host-asset preflight"; \
 	fi
-	@if [ -f $(MICROVM_OUT_DIR)/vmlinuz ]; then \
-		$(MAKE) install-microvm; \
-	else \
-		echo "[deploy] microvm artifacts absent — run 'make install-microvm' for first-time setup"; \
-	fi
 	sudo systemctl daemon-reload
 	sudo systemctl restart spinifex.target

@@ -367,16 +419,25 @@ clean:
 	rm -f ./bin/$(GO_PROJECT_NAME)
 	rm -rf spinifex/runtime/roles/spinifexui/frontend/dist

+# On dnf hosts the runtime packages and repositories come from setup.sh, so the
+# two cannot drift; nbdkit-devel, for building the plugin, is only in CodeReady.
 install-system:
 	@echo -e "\n....Installing system dependencies for $(ARCH)...."
-	sudo apt-get update && sudo DEBIAN_FRONTEND=noninteractive apt-get install -y \
-		-o Dpkg::Options::="--force-confdef" \
-		-o Dpkg::Options::="--force-confold" \
-		nbdkit nbdkit-plugin-dev pkg-config qemu-system-x86 qemu-system-arm qemu-utils \
-		ovmf qemu-efi-aarch64 \
-		libvirt-daemon-system libvirt-clients libvirt-dev make gcc jq curl \
-		iproute2 netcat-openbsd openssh-client wget git unzip sudo xz-utils file \
-		ovn-central ovn-host openvswitch-switch dhcpcd-base
+	@pm=$$(INSTALL_SPINIFEX_LIB_ONLY=1 bash -c '. scripts/setup.sh && distro_pkg_manager') || exit 1; \
+	if [ "$$pm" = dnf ]; then \
+		INSTALL_SPINIFEX_LIB_ONLY=1 bash -c '. scripts/setup.sh && setup_sudo && detect_os && install_dnf_deps' && \
+		sudo dnf config-manager --set-enabled ol10_codeready_builder && \
+		sudo dnf install -y nbdkit-devel pkgconf-pkg-config make gcc openssh-clients git sudo; \
+	else \
+		sudo apt-get update && sudo DEBIAN_FRONTEND=noninteractive apt-get install -y \
+			-o Dpkg::Options::="--force-confdef" \
+			-o Dpkg::Options::="--force-confold" \
+			nbdkit nbdkit-plugin-dev pkg-config qemu-system-x86 qemu-system-arm qemu-utils \
+			ovmf qemu-efi-aarch64 \
+			libvirt-daemon-system libvirt-clients libvirt-dev make gcc jq curl \
+			iproute2 netcat-openbsd openssh-client wget git unzip sudo xz-utils file \
+			ovn-central ovn-host openvswitch-switch dhcpcd-base; \
+	fi

 install-go:
 	@echo -e "\n....Installing Go 1.27.3 for $(ARCH) ($(GO_ARCH))...."
@@ -438,14 +499,17 @@ distro-amd64:
 		-f build/Dockerfile.distro \
 		--output type=local,dest=dist/amd64/ \
 		../
-	@if [ -f $(MICROVM_OUT_DIR)/vmlinuz ] && [ -f $(MICROVM_OUT_DIR)/initramfs.cpio.gz ]; then \
+	@if [ -f $(MICROVM_OUT_DIR)/vmlinuz ] && [ -f $(MICROVM_OUT_DIR)/initramfs.cpio.gz ] && [ -f $(MICROVM_OUT_DIR)/qboot.rom ] && [ -f $(QBOOT_SOURCE) ]; then \
 		echo "[distro-amd64] staging microvm artifacts into tarball"; \
 		mkdir -p dist/amd64/microvm; \
-		cp $(MICROVM_OUT_DIR)/vmlinuz $(MICROVM_OUT_DIR)/initramfs.cpio.gz dist/amd64/microvm/; \
+		cp $(MICROVM_OUT_DIR)/vmlinuz $(MICROVM_OUT_DIR)/initramfs.cpio.gz $(MICROVM_OUT_DIR)/qboot.rom $(MICROVM_OUT_DIR)/qboot.LICENSE dist/amd64/microvm/; \
+		cp $(QBOOT_SOURCE) dist/; \
 	else \
 		echo "[distro-amd64] WARNING: microvm artifacts missing — tarball will not include them"; \
 		echo "[distro-amd64]          run 'make build-microvm-image' before 'make distro-amd64'"; \
 	fi
+	./scripts/lib/stage-sqs-store-microvm.sh $(SQS_STORE_MICROVM_OUT_DIR) dist/amd64
+	./scripts/lib/stage-sqs-store-microvm.sh $(SQS_WORKER_MICROVM_OUT_DIR) dist/amd64 sqs-worker
 	tar -czf dist/spinifex-$(VERSION)-linux-amd64.tar.gz -C dist/amd64 .
 	sha256sum dist/spinifex-$(VERSION)-linux-amd64.tar.gz > dist/spinifex-$(VERSION)-linux-amd64.tar.gz.sha256

@@ -464,7 +528,7 @@ distro-arm64:
 distro-clean:
 	rm -rf dist/

-.PHONY: test-package-check build build-ui build-installer build-lb-agent build-ecs-agent build-system-image build-eks-node-image import-eks-node-image publish-eks-node-image build-ecs-node-image import-ecs-node-image build-rds-postgres-image import-rds-postgres-image build-rds-mariadb-image import-rds-mariadb-image build-microvm-image install-microvm go_build preflight test test-cover test-race diff-coverage bench fuzz test-actions test-images test-build-scripts test-harness test-integration generate-aws-model-coverage aws-model-coverage test-segscan-oracle manifest-check manifest-lint manifest-lint-update \
+.PHONY: test-package-check build build-ui test-ui-e2e ui-e2e-container ui-baselines build-installer build-lb-agent build-ecs-agent build-system-image build-eks-node-image import-eks-node-image publish-eks-node-image build-ecs-node-image import-ecs-node-image build-rds-postgres-image import-rds-postgres-image build-rds-mariadb-image import-rds-mariadb-image build-microvm-image build-sqs-store-microvm-image build-sqs-worker-microvm-image install-microvm install-sqs-store-microvm install-sqs-worker-microvm go_build preflight test test-cover test-race diff-coverage bench fuzz test-actions test-images test-build-scripts test-harness test-integration generate-aws-model-coverage aws-model-coverage test-segscan-oracle \
 	deploy reinstall clean \
 	install-system install-go install-aws quickinstall \
 	lint fix govulncheck nilaway \
@@ -54,7 +54,7 @@ For Terraform, start from the [example workbooks](https://docs.mulgadc.com/docs/

 ## Services

-The platform serves **392 AWS API operations** across the services below. See exactly which ones in the [coverage report](https://docs.mulgadc.com/coverage).
+The platform serves **405 AWS API operations** across the services below. See exactly which ones in the [coverage report](https://docs.mulgadc.com/coverage).

 | Service | What it is | Built on | Status |
 | --- | --- | --- | --- |
@@ -68,8 +68,10 @@ The platform serves **392 AWS API operations** across the services below. See ex
 | **ECR** | Container registry | Predastore, deduplicated per account | Available |
 | **ECS** | Container service | Spinifex instances | Available |
 | **RDS** | Databases | PostgreSQL and MariaDB, with automated backups | Available |
+| **Route 53** | DNS | [Northstar](https://github.com/mulgadc/northstar): authoritative DNS with hosted zones | Available |
 | **Bedrock** | AI deployment | vLLM for text generation and Hugging Face TEI for embeddings and reranking | Q4 2026 |
 | **Lambda** | Serverless functions | | Q1 2027 |
+| **SQS** | Message queues | | Q1 2027 |

 Roadmap services ship under the same AWS API surface, so code written for AWS today works the moment they land. [Track what's shipped in the release notes](https://github.com/mulgadc/spinifex/releases).

@@ -29,8 +29,16 @@ var allowlist = []allowance{

 	{From: "agents/rds/agent", To: "domains/rds/instance", Debt: "agents/rds/agent (dependency, narrowed)", Reason: "command protocol, wire and record types; needs a guest/controller wire contract"},

-	{From: "cmd/aws-model-coverage", To: "domains/acm/awsapi", Debt: "cmd/aws-model-coverage", Reason: "reads declared ECR/ACM inventories; named exception, must not construct a registration"},
-	{From: "cmd/aws-model-coverage", To: "domains/ecr/awsapi", Debt: "cmd/aws-model-coverage", Reason: "reads declared ECR/ACM inventories; named exception, must not construct a registration"},
+	{From: "cmd/aws-model-coverage", To: "domains/acm/awsapi", Debt: "cmd/aws-model-coverage", Reason: "reads declared service inventories; named exception, must not construct a registration"},
+	{From: "cmd/aws-model-coverage", To: "domains/ecr/awsapi", Debt: "cmd/aws-model-coverage", Reason: "reads declared service inventories; named exception, must not construct a registration"},
+	{From: "cmd/aws-model-coverage", To: "domains/route53/awsapi", Debt: "cmd/aws-model-coverage", Reason: "reads declared service inventories; named exception, must not construct a registration"},
+	{From: "cmd/aws-model-coverage", To: "domains/elbv2/awsapi", Debt: "cmd/aws-model-coverage", Reason: "reads declared service inventories; named exception, must not construct a registration"},
+	{From: "cmd/aws-model-coverage", To: "domains/rds/awsapi", Debt: "cmd/aws-model-coverage", Reason: "reads declared service inventories; named exception, must not construct a registration"},
+	{From: "cmd/aws-model-coverage", To: "domains/ec2/awsapi", Debt: "cmd/aws-model-coverage", Reason: "reads declared service inventories; named exception, must not construct a registration"},
+	{From: "cmd/aws-model-coverage", To: "domains/ecs/awsapi", Debt: "cmd/aws-model-coverage", Reason: "reads declared service inventories; named exception, must not construct a registration"},
+	{From: "cmd/aws-model-coverage", To: "domains/eks/awsapi", Debt: "cmd/aws-model-coverage", Reason: "reads declared service inventories; named exception, must not construct a registration"},
+	{From: "cmd/aws-model-coverage", To: "domains/iam/awsapi", Debt: "cmd/aws-model-coverage", Reason: "reads declared service inventories; named exception, must not construct a registration"},
+	{From: "cmd/aws-model-coverage", To: "domains/sts/awsapi", Debt: "cmd/aws-model-coverage", Reason: "reads declared service inventories; named exception, must not construct a registration"},

 	{From: "runtime/daemon", To: "domains/ec2/awsapi/instance", Debt: "Instance adapter consumers outside EC2", Reason: "runtime/daemon/eks_cp_control.go and rds_deps.go call the instance adapter; needs an EC2 capability"},

@@ -52,9 +60,6 @@ var allowlist = []allowance{
 	{From: "domains/ec2/awsapi/instance", To: "runtime/compute/cache", Debt: "domains/ec2/awsapi/instance (dependency)", Reason: "node liveness and VM state; needs an EC2-owned projection"},
 	{From: "domains/ec2/awsapi/instance", To: "runtime/compute/vm", Debt: "domains/ec2/awsapi/instance (dependency)", Reason: "node liveness and VM state; needs an EC2-owned projection"},

-	{From: "domains/ec2/awsapi/spotinstance", To: "domains/admission/quota", Debt: "domains/ec2/awsapi/spotinstance (dependency)", Reason: "RequestSpotInstances.go uses IAMService and quota.Service"},
-	{From: "domains/ec2/awsapi/spotinstance", To: "domains/iam", Debt: "domains/ec2/awsapi/spotinstance (dependency)", Reason: "RequestSpotInstances.go uses IAMService and quota.Service"},
-
 	{From: "domains/ec2/eip", To: "domains/network/topology", Debt: "EC2 → network implementation", Reason: "EC2 reads network implementation; clears with the EC2-owned projection"},

 	{From: "domains/ec2/guestmetadata", To: "domains/dns", Debt: "Guest metadata: the cross-domain and runtime imports above", Reason: "guest-metadata relocation residual; needs capabilities, projections or contracts"},
@@ -91,42 +96,15 @@ var allowlist = []allowance{
 	{From: "domains/ochre", To: "domains/ec2/systeminstance", Debt: "Cross-domain", Reason: "consumes another domain's implementation rather than a capability"},
 	{From: "domains/ochre", To: "domains/iam", Debt: "IAM implementation consumers in STS, Ochre and ELBv2", Reason: "secret encrypt/decrypt helpers; needs a shared crypto boundary"},
 	{From: "domains/ochre", To: "domains/network/systemvpc", Debt: "Cross-domain", Reason: "consumes another domain's implementation rather than a capability"},
-	{From: "domains/ochre", To: "ingress/aws/bodyscope", Debt: "domains/ochre (action adapters)", Reason: "Bedrock action adapters still live in the Ochre package; they move to domains/ochre/awsapi"},
 	{From: "domains/ochre", To: "runtime/compute/gpu", Debt: "Domain → runtime", Reason: "GPU runtime state read directly; needs a domain-owned projection"},

 	{From: "domains/ochre/vector", To: "domains/iam", Debt: "Cross-domain", Reason: "consumes another domain's implementation rather than a capability"},

-	{From: "gateway", To: "domains/ec2/awsapi", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/account", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/capacityreservation", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/eigw", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/eip", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/idem", Debt: "gateway/ec2.go (registration; ec2_idempotency.go)", Reason: "dispatch-time idempotency wrapper; moves with the EC2 dispatch map"},
-	{From: "gateway", To: "domains/ec2/awsapi/igw", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/image", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/instance", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/key", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/launchtemplate", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/natgw", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/placementgroup", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/routetable", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/snapshot", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/spotinstance", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/tags", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/volume", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/vpc", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ec2/awsapi/zone", Debt: "gateway/ec2.go (registration)", Reason: "EC2 dispatch map stays in gateway until the ingress registration seam"},
-
 	{From: "domains/rds/awsapi", To: "domains/ec2/awsapi/instance", Debt: "Instance adapter consumers outside EC2", Reason: "enginecatalog.go calls the instance adapter; needs an EC2 capability"},

 	{From: "domains/sts/awsapi", To: "domains/iam", Debt: "IAM implementation consumers in STS, Ochre and ELBv2", Reason: "GetCallerIdentity takes IAMService; needs an IAM capability"},

-	{From: "gateway", To: "domains/eks/awsapi", Debt: "gateway (service registration)", Reason: "IAM, STS, EKS and RDS dispatch stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/iam/awsapi", Debt: "gateway (service registration)", Reason: "IAM, STS, EKS and RDS dispatch stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/rds/awsapi", Debt: "gateway (service registration)", Reason: "IAM, STS, EKS and RDS dispatch stays in gateway until the ingress registration seam"},
 	{From: "gateway", To: "domains/sts/awsapi", Debt: "gateway (service registration)", Reason: "IAM, STS, EKS and RDS dispatch stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/ecs/awsapi", Debt: "gateway (service registration)", Reason: "ECS and ELBv2 dispatch stays in gateway until the ingress registration seam"},
-	{From: "gateway", To: "domains/elbv2/awsapi", Debt: "gateway (service registration)", Reason: "ECS and ELBv2 dispatch stays in gateway until the ingress registration seam"},
 	{From: "domains/ecs", To: "domains/ec2/ebs/metadata", Debt: "close-out: handlers/ecs (ADR-0004 blocked-dependency list; inventory-ecs boundary imports)", Reason: "legacy ECS calls another domain's implementation as an internal API"},
 	{From: "domains/ecs", To: "domains/ec2/eip", Debt: "close-out: handlers/ecs (ADR-0004 blocked-dependency list; inventory-ecs boundary imports)", Reason: "legacy ECS calls another domain's implementation as an internal API"},
 	{From: "domains/ecs", To: "domains/ec2/instancetypes", Debt: "close-out: handlers/ecs (ADR-0004 blocked-dependency list; inventory-ecs boundary imports)", Reason: "legacy ECS calls another domain's implementation as an internal API"},
@@ -159,6 +137,4 @@ var allowlist = []allowance{
 	{From: "domains/rds/instance", To: "runtime/compute/vm", Debt: "close-out: handlers/rds (ADR-0004 blocked-dependency list; inventory-rds boundary imports)", Reason: "legacy RDS calls another domain's or runtime implementation directly"},

 	{From: "domains/sts", To: "domains/iam", Debt: "IAM implementation consumers in STS, Ochre and ELBv2", Reason: "STS reads IAM service, trust-policy and OIDC-provider state; needs an IAM capability"},
-
-	{From: "operator/cli", To: "gateway", Debt: "operator/cli (dependency)", Reason: "command tree imports implementations; needs public capabilities"},
 }
@@ -213,3 +213,28 @@ func TestADR0001_InterimDaemonDirectionCoversRuntimeDaemon(t *testing.T) {
 		t.Error("gateway -> runtime/host/process is rejected, but it is not the daemon")
 	}
 }
+
+// TestADR0002_S7_RequestRecordRejectsOutsiders: the request record is readable
+// by the gateway and ingress/aws only.
+func TestADR0002_S7_RequestRecordRejectsOutsiders(t *testing.T) {
+	record, ok := classify("spinifex/ingress/aws/reqctx")
+	if !ok || !record.requestRecord {
+		t.Fatalf("ingress/aws/reqctx is not classified as the request record: %+v", record)
+	}
+	r := ruleByID("S7_RequestRecordPrivateToIngress")
+	for name, wantRejected := range map[string]bool{
+		"domains/ecs/awsapi":   true,
+		"runtime/roles/awsgw":  true,
+		"operator/admin":       true,
+		"gateway":              false,
+		"ingress/aws/dispatch": false,
+	} {
+		from, ok := classify(fullPath(name))
+		if !ok {
+			t.Fatalf("%s is unclassified", name)
+		}
+		if got := r.forbids(from, record, name); got != wantRejected {
+			t.Errorf("%s -> ingress/aws/reqctx rejected=%v, want %v", name, got, wantRejected)
+		}
+	}
+}
@@ -43,6 +43,8 @@ type class struct {
 	domain string // owning domain for domain-shaped and agent packages
 	kind   domainKind
 	daemon bool // runtime/daemon, the interim daemon -> gateway -> handlers root
+	// requestRecord marks ingress/aws/reqctx, the request record ADR-0002:S7 keeps private to ingress.
+	requestRecord bool
 }

 func (c class) domainShaped() bool { return c.kind != notDomain }
@@ -103,7 +105,7 @@ func classifyApp(app []string) (class, bool) {
 	case "foundation":
 		return class{layer: layerFoundation}, true
 	case "ingress":
-		return class{layer: layerIngress}, true
+		return class{layer: layerIngress, requestRecord: len(app) >= 3 && app[1] == "aws" && app[2] == "reqctx"}, true
 	case "runtime":
 		return class{layer: layerRuntime, daemon: len(app) >= 2 && app[1] == "daemon"}, true
 	case "operator":
@@ -11,7 +11,6 @@ import (
 // completed move whose line must be deleted.
 var frozenLegacyPackages = []string{
 	"gateway",
-	"gateway/spx",
 }

 // retiredLegacyRoots were legacy roots whose every package has moved to its
@@ -1,6 +1,9 @@
 package layering

-import "slices"
+import (
+	"slices"
+	"strings"
+)

 // rule is one ADR-0001 dependency rule over a classified import edge.
 type rule struct {
@@ -151,6 +154,18 @@ var rules = []rule{
 				(importsDomainImplementation(t) || t.layer == layerAgents)
 		},
 	},
+	{
+		id:      "S7_RequestRecordPrivateToIngress",
+		clause:  "ADR-0002:S7: the request record is private to ingress, so no domain, runtime or operator package reads it; only the gateway and packages under ingress/aws may import ingress/aws/reqctx",
+		hint:    "receive what you need through dispatch.Invocation instead of the request record",
+		subject: func(_ class, _ string) bool { return true },
+		forbids: func(_, t class, fromName string) bool {
+			if !t.requestRecord {
+				return false
+			}
+			return fromName != "gateway" && !strings.HasPrefix(fromName, "ingress/aws/")
+		},
+	},
 	{
 		id:      "Interim_DaemonGatewayHandlersDirection",
 		clause:  "during migration the existing daemon -> gateway -> handlers rule remains the guardrail (Q-20, Q-29)",
@@ -42,9 +42,16 @@ RUN COMMIT=$(cd spinifex && git rev-parse --short HEAD 2>/dev/null || echo "unkn
 # Build nbdkit viperblock plugin (.so)
 RUN cd viperblock && GOFIPS140=v1.0.0 go build -o /out/nbdkit-viperblock-plugin.so -buildmode=c-shared nbd/viperblock.go

+# One tarball ships to every OS. The ceiling is what both binaries need today,
+# so a rise is a decision rather than an accident; every supported OS has 2.39+.
+RUN spinifex/scripts/check-glibc-ceiling.sh 2.34 /out/spx /out/nbdkit-viperblock-plugin.so
+
 # Copy setup scripts and systemd units
 RUN cp spinifex/scripts/setup.sh /out/setup.sh 2>/dev/null || true
 RUN cp spinifex/scripts/setup-ovn.sh /out/setup-ovn.sh
+RUN cd spinifex/scripts && cp install-node.sh smoke-test.sh node-reset.sh uninstall-spx.sh /out/
+RUN mkdir -p /out/udev && cp spinifex/build/udev/* /out/udev/
+RUN mkdir -p /out/lib && cp spinifex/scripts/lib/distro.sh /out/lib/distro.sh
 RUN mkdir -p /out/systemd && cp spinifex/build/systemd/* /out/systemd/
 RUN mkdir -p /out/scripts && cp spinifex/build/scripts/* /out/scripts/
 RUN cp spinifex/build/logrotate/spinifex /out/logrotate-spinifex
@@ -0,0 +1,96 @@
+#!/bin/sh
+mount -t proc  proc  /proc
+mount -t sysfs sys   /sys
+mount -t devtmpfs dev /dev
+
+# virtio_mmio is built-in (CONFIG_VIRTIO_MMIO=y) so virtio-net devices are
+# already enumerated by the time PID 1 runs — no modprobe needed.
+# qemu_fw_cfg and virtio_net are =m in stock Alpine linux-virt, so load them.
+#
+# modprobe needs a usable modules.dep; when the image is built on a host whose
+# depmod can't produce one, busybox modprobe silently no-ops and the module
+# never loads — fw_cfg then never appears and init panics. insmod by explicit
+# path needs no modules.dep, so fall back to it. Handles .ko and .ko.gz.
+load_mod() {
+    modprobe "$1" 2>/dev/null && return 0
+    for ko in $(find /lib/modules -name "$1.ko" -o -name "$1.ko.gz" 2>/dev/null); do
+        insmod "$ko" 2>/dev/null && return 0
+        case "$ko" in
+            *.gz) gunzip -c "$ko" >/tmp/mod.ko 2>/dev/null \
+                  && insmod /tmp/mod.ko 2>/dev/null && return 0 ;;
+        esac
+    done
+    return 1
+}
+load_mod qemu_fw_cfg || load_mod fw_cfg_sysfs || true
+# virtio_net depends on net_failover, which depends on failover. The load_mod
+# insmod fallback loads a single .ko with no dependency resolution, and
+# modules.dep is absent when the image is built on a kmod-less host (depmod
+# skipped), so modprobe can't resolve the chain either. Load it explicitly
+# bottom-up so the NIC comes up regardless of how the image was built —
+# without this the agent has no network and the LB hangs in provisioning.
+load_mod failover || true
+load_mod net_failover || true
+load_mod virtio_net || true
+
+NETCFG=/sys/firmware/qemu_fw_cfg/by_name/opt/spinifex/netcfg/raw
+i=0
+while [ "$i" -lt 25 ]; do
+    [ -f "$NETCFG" ] && break
+    sleep 0.04
+    i=$((i + 1))
+done
+
+if [ ! -f "$NETCFG" ]; then
+    # Dump host/guest fw_cfg state to the serial console before init exits so a
+    # "netcfg not available" panic is diagnosable from the captured console log
+    # (module-not-loaded vs qemu-not-exposing-by_name vs wrong machine type).
+    echo "[init] FATAL: fw_cfg netcfg not available" >&2
+    echo "[init] dmi: vendor=$(cat /sys/class/dmi/id/sys_vendor 2>/dev/null) product=$(cat /sys/class/dmi/id/product_name 2>/dev/null) bios=$(cat /sys/class/dmi/id/bios_version 2>/dev/null)" >&2
+    echo "[init] qemu_fw_cfg dir present: $([ -d /sys/firmware/qemu_fw_cfg ] && echo yes || echo NO)" >&2
+    echo "[init] modules: $(lsmod 2>/dev/null | awk 'NR>1{print $1}' | tr '\n' ',')" >&2
+    echo "[init] by_name listing:" >&2
+    ls -la /sys/firmware/qemu_fw_cfg/by_name/ 2>&1 | sed 's/^/[init]   /' >&2
+    ls -la /sys/firmware/qemu_fw_cfg/by_name/opt/ 2>&1 | sed 's/^/[init]   /' >&2
+    echo "[init] dmesg fw_cfg:" >&2
+    dmesg 2>/dev/null | grep -i fw_cfg | sed 's/^/[init]   /' >&2
+    exit 1
+fi
+
+. "$NETCFG"
+
+for n in 0 1 2 3 4 5; do
+    eval mac=\$NIC${n}_MAC
+    [ -z "$mac" ] && continue
+
+    iface=""
+    for attempt in 1 2 3 4 5; do
+        iface=$(ip -o link | awk -v m="$mac" 'tolower($0) ~ tolower(m) {print $2}' | tr -d :)
+        [ -n "$iface" ] && break
+        sleep 0.02
+    done
+
+    if [ -z "$iface" ]; then
+        echo "[init] WARNING: no interface found for MAC $mac (NIC$n), skipping" >&2
+        continue
+    fi
+
+    eval cidr=\$NIC${n}_CIDR
+    eval gw=\$NIC${n}_GW
+    eval default=\$NIC${n}_DEFAULT
+    eval rdst=\$NIC${n}_ROUTE_DST
+    eval rvia=\$NIC${n}_ROUTE_VIA
+
+    ip link set "$iface" up
+    [ -n "$cidr" ] && ip addr add "$cidr" dev "$iface"
+    [ "$default" = "1" ] && [ -n "$gw" ] && ip route add default via "$gw" dev "$iface"
+    [ -n "$rdst" ] && [ -n "$rvia" ] && ip route add "$rdst" via "$rvia" dev "$iface"
+done
+
+# The store image takes no lb-agent environment. A CA bundle is optional here:
+# nothing in this image dials out yet.
+CA_CERT=/sys/firmware/qemu_fw_cfg/by_name/opt/spinifex/ca-cert/raw
+mkdir -p /etc/ssl/certs
+[ -f "$CA_CERT" ] && cp "$CA_CERT" /etc/ssl/certs/ca-certificates.crt
+
+exec /sbin/init
@@ -0,0 +1,10 @@
+# The power button loads here, not in /init: busybox init has its SIGUSR2
+# (poweroff) handler installed before sysinit runs, so a press can never reach
+# the pre-exec shell PID 1. A press before this line is lost; the host re-sends.
+::sysinit:/sbin/modprobe tiny_power_button power_signal=12
+::sysinit:/usr/local/lib/sqs-store/sqs-store-prepare
+::respawn:/usr/local/lib/sqs-store/sqs-store-run
+::ctrlaltdel:/sbin/reboot
+# One action: the store stops and its disk unmounts first, and the generic
+# umount -a -r runs only if the store is gone.
+::shutdown:/usr/local/lib/sqs-store/sqs-store-stop
@@ -0,0 +1,216 @@
+# sqs-store-lib.sh — shared by the SQS store guest's prepare, run and stop
+# steps. Sourced, never executed. Every path and command is overridable so the
+# host test can drive the same code against a fake fw_cfg, sysfs and tools.
+#
+# Guest contract, read from fw_cfg under opt/spinifex/sqs-store/:
+#   data-serial  virtio-blk serial of the already-formatted ext4 data disk
+#   listen-addr  IPv4 address on a configured guest NIC; the only listener
+#   client-nkey  NATS user NKey public key allowed to connect; never a seed
+# Anything absent, malformed or ambiguous refuses the store. Nothing formats.
+
+SQS_FWCFG=${SQS_FWCFG:-/sys/firmware/qemu_fw_cfg/by_name/opt/spinifex/sqs-store}
+SQS_SYSBLOCK=${SQS_SYSBLOCK:-/sys/block}
+SQS_DEVDIR=${SQS_DEVDIR:-/dev}
+# The host test cannot create block nodes, so it relaxes this to -e.
+SQS_NODE_TEST=${SQS_NODE_TEST:--b}
+SQS_RUN=${SQS_RUN:-/run/sqs-store}
+SQS_DATA=${SQS_DATA:-/data}
+SQS_CONSOLE=${SQS_CONSOLE:-/dev/console}
+SQS_MOUNTS=${SQS_MOUNTS:-/proc/mounts}
+SQS_PORT=${SQS_PORT:-4222}
+SQS_STOP_BUDGET_TENTHS=${SQS_STOP_BUDGET_TENTHS:-120}
+NATS_BIN=${NATS_BIN:-/usr/local/bin/nats-server}
+E2FSCK=${E2FSCK:-e2fsck}
+MOUNT=${MOUNT:-mount}
+UMOUNT=${UMOUNT:-umount}
+IP=${IP:-ip}
+
+sqs_log() { echo "[sqs-store] $*" >> "$SQS_CONSOLE"; }
+sqs_fatal() { sqs_log "FATAL: $*"; return 1; }
+
+# sqs_fwcfg <name>: print one fw_cfg value, refusing a missing, empty, oversized
+# or multi-token file. The value is not echoed on failure.
+sqs_fwcfg() {
+    f="$SQS_FWCFG/$1/raw"
+    [ -f "$f" ] || { sqs_fatal "fw_cfg $1 not supplied"; return 1; }
+    [ "$(wc -c < "$f")" -le 128 ] || { sqs_fatal "fw_cfg $1 is oversized"; return 1; }
+    v=$(cat "$f")
+    case "$v" in
+        "" | *[[:space:]]*) sqs_fatal "fw_cfg $1 must be exactly one value"; return 1 ;;
+    esac
+    printf '%s' "$v"
+}
+
+# A volume serial is the volume ID with dashes stripped: vol + 17 hex.
+sqs_valid_serial() {
+    [ "${#1}" -eq 20 ] || return 1
+    case "$1" in vol*) ;; *) return 1 ;; esac
+    case "${1#vol}" in *[!0-9a-f]*) return 1 ;; esac
+}
+
+# A NATS user public NKey: "U" then 55 base32 characters. The checksum is
+# left to the pinned server's own config test.
+sqs_valid_user_nkey() {
+    [ "${#1}" -eq 56 ] || return 1
+    case "$1" in U*) ;; *) return 1 ;; esac
+    case "$1" in *[!A-Z2-7]*) return 1 ;; esac
+}
+
+sqs_valid_ipv4() {
+    case "$1" in *[!0-9.]* | .* | *. | *..*) return 1 ;; esac
+    old_ifs=$IFS; IFS=.
+    # shellcheck disable=SC2086 # split on dots on purpose
+    set -- $1
+    IFS=$old_ifs
+    [ "$#" -eq 4 ] || return 1
+    for o in "$@"; do
+        [ "${#o}" -le 3 ] && [ "$o" -le 255 ] || return 1
+        case "$o" in 0?*) return 1 ;; esac
+    done
+}
+
+# sqs_find_disk <serial>: print the /dev node of the one block device whose
+# sysfs serial matches exactly. Zero or several matches refuse.
+sqs_find_disk() {
+    found="" n=0
+    for b in "$SQS_SYSBLOCK"/*; do
+        [ -r "$b/serial" ] || continue
+        [ "$(cat "$b/serial")" = "$1" ] || continue
+        found=${b##*/}
+        n=$((n + 1))
+    done
+    [ "$n" -eq 1 ] || { sqs_fatal "expected one disk with serial $1, found $n"; return 1; }
+    test "$SQS_NODE_TEST" "$SQS_DEVDIR/$found" || { sqs_fatal "disk $found has no block device node"; return 1; }
+    printf '%s' "$SQS_DEVDIR/$found"
+}
+
+# sqs_has_addr <ipv4>: the address is assigned to a non-loopback interface.
+sqs_has_addr() {
+    "$IP" -o -4 addr show 2>/dev/null \
+        | awk -v a="$1" '$2 != "lo" { split($4, p, "/"); if (p[1] == a) f = 1 } END { exit !f }'
+}
+
+# sqs_alive <pid>: running or stopped, not exited (a zombie counts as exited).
+sqs_alive() { [ -r "/proc/$1/status" ] && ! grep -q '^State:[[:space:]]*Z' "/proc/$1/status" 2>/dev/null; }
+
+# sqs_starttime <pid>: kernel start time (stat field 22). It survives exec, and
+# pairing it with the pid reduces PID-reuse risk; without a pidfd the check and
+# the signal stay separate steps, so it is not an absolute guarantee.
+sqs_starttime() { sed 's/.*) //' "/proc/$1/stat" 2>/dev/null | cut -d' ' -f20; }
+
+# sqs_same <pid> <starttime>: the recorded process is still alive.
+sqs_same() { [ -n "$2" ] && sqs_alive "$1" && [ "$(sqs_starttime "$1")" = "$2" ]; }
+
+# sqs_prepare: validate the contract, check and mount the data disk, write the
+# server config and mark the store ready. Returns non-zero, unready, on any refusal.
+sqs_prepare() {
+    rm -f "$SQS_RUN/ready"
+    mkdir -p "$SQS_RUN"
+    serial=$(sqs_fwcfg data-serial) || return 1
+    sqs_valid_serial "$serial" || { sqs_fatal "data-serial is not a volume serial"; return 1; }
+    addr=$(sqs_fwcfg listen-addr) || return 1
+    sqs_valid_ipv4 "$addr" || { sqs_fatal "listen-addr is not an IPv4 address"; return 1; }
+    nkey=$(sqs_fwcfg client-nkey) || return 1
+    sqs_valid_user_nkey "$nkey" || { sqs_fatal "client-nkey is not a user public NKey"; return 1; }
+    sqs_has_addr "$addr" || { sqs_fatal "listen-addr $addr is not on any guest NIC"; return 1; }
+
+    # The server checks the config, including the NKey checksum, before the
+    # disk is touched.
+    umask 077
+    cat > "$SQS_RUN/nats.conf" <<EOF
+listen: "$addr:$SQS_PORT"
+server_name: "sqs-store"
+jetstream {
+  store_dir: "$SQS_DATA"
+  sync_interval: always
+}
+authorization {
+  users = [ { nkey: "$nkey" } ]
+}
+EOF
+    if ! "$NATS_BIN" -t -c "$SQS_RUN/nats.conf" > "$SQS_RUN/config-test.out" 2>&1; then
+        sqs_fatal "server config test failed: $(sed 's/U[A-Z2-7]\{55\}/<nkey>/g' "$SQS_RUN/config-test.out" | tr '\n' ' ')"
+        return 1
+    fi
+
+    dev=$(sqs_find_disk "$serial") || return 1
+    rc=0
+    "$E2FSCK" -p "$dev" > "$SQS_RUN/fsck.out" 2>&1 || rc=$?
+    sqs_log "e2fsck -p $dev rc=$rc"
+    # 0 clean, 1 corrected: proceed. Anything higher, including a disk that is
+    # not ext4 at all, refuses; the guest never formats.
+    [ "$rc" -le 1 ] || { sqs_fatal "e2fsck refused $dev: $(tr '\n' ' ' < "$SQS_RUN/fsck.out")"; return 1; }
+    mkdir -p "$SQS_DATA"
+    "$MOUNT" -t ext4 -o nodev,nosuid,noexec "$dev" "$SQS_DATA" || { sqs_fatal "mount $dev failed"; return 1; }
+    : > "$SQS_RUN/ready"
+    sqs_log "prepared: $dev mounted at $SQS_DATA, config written for $addr:$SQS_PORT; server not started yet"
+}
+
+# sqs_run: the respawned supervisor body. Its identity is recorded before the
+# stopping check: a stop that misses it set stopping first, so this parks; a
+# stop that reads it signals this process whether or not it has exec'd yet.
+sqs_run() {
+    mkdir -p "$SQS_RUN"
+    echo "$$ $(sqs_starttime $$)" > "$SQS_RUN/pid"
+    if [ -e "$SQS_RUN/stopping" ]; then
+        sqs_log "run: shutdown in progress, not starting"
+        return 2
+    fi
+    if [ ! -e "$SQS_RUN/ready" ]; then
+        sqs_log "run: store not prepared, not starting"
+        return 2
+    fi
+    sqs_log "run: starting pid=$$"
+    # ADR-0011:S11 bounds the store's Go heap; the guest must be larger.
+    GOMEMLIMIT=512MiB exec "$NATS_BIN" -c "$SQS_RUN/nats.conf"
+}
+
+# sqs_stop: first shutdown action. Block any respawn, SIGTERM the recorded run
+# process (a server, or a start not yet exec'd), SIGKILL only after the budget,
+# then sync and unmount. Each signal rechecks the identity; a survivor keeps the disk mounted.
+sqs_stop() {
+    mkdir -p "$SQS_RUN"
+    : > "$SQS_RUN/stopping"
+    sqs_log "stop: begin"
+    rec=$(cat "$SQS_RUN/pid" 2>/dev/null || true)
+    pid=${rec%% *} st=${rec#* }
+    [ "$st" != "$rec" ] || st=""
+    if [ -n "$pid" ] && sqs_same "$pid" "$st"; then
+        kill -TERM "$pid" 2>/dev/null || true
+        sqs_log "stop: SIGTERM sent to run pid=$pid"
+        i=0
+        while sqs_same "$pid" "$st" && [ "$i" -lt "$SQS_STOP_BUDGET_TENTHS" ]; do sleep 0.1; i=$((i + 1)); done
+        if sqs_same "$pid" "$st"; then
+            sqs_log "stop: run pid=$pid still running after $((SQS_STOP_BUDGET_TENTHS / 10))s, SIGKILL"
+            kill -KILL "$pid" 2>/dev/null || true
+            j=0
+            while sqs_same "$pid" "$st" && [ "$j" -lt 30 ]; do sleep 0.1; j=$((j + 1)); done
+        fi
+        if sqs_same "$pid" "$st"; then
+            sqs_log "stop: run pid=$pid STILL ALIVE, leaving $SQS_DATA mounted"
+            return 1
+        fi
+        sqs_log "stop: store exited after $((i / 10)).$((i % 10))s"
+    else
+        sqs_log "stop: no running store"
+    fi
+    sync
+    if grep -q " $SQS_DATA " "$SQS_MOUNTS" 2>/dev/null; then
+        rc=0
+        "$UMOUNT" "$SQS_DATA" || rc=$?
+        sqs_log "stop: umount $SQS_DATA rc=$rc"
+    fi
+}
+
+# sqs_shutdown: the only shutdown action. The generic read-only remount runs
+# only after the store stopped; a surviving writer keeps every mount as it is.
+sqs_shutdown() {
+    if ! sqs_stop; then
+        sync
+        sqs_log "shutdown: store survived, skipping umount -a -r"
+        return 1
+    fi
+    rc=0
+    "$UMOUNT" -a -r || rc=$?
+    echo "[init] shutdown: umount -a -r rc=$rc" >> "$SQS_CONSOLE"
+}
@@ -0,0 +1,9 @@
+#!/bin/sh
+# sysinit: load the disk drivers, then validate the guest contract, check and
+# mount the data disk and write the server config. A refusal leaves the store
+# unprepared, so the run step never starts it.
+. /usr/local/lib/sqs-store/sqs-store-lib.sh
+modprobe virtio_blk 2>/dev/null || true
+modprobe ext4 2>/dev/null || true
+sqs_prepare || sqs_log "store refused; it will not start"
+exit 0
@@ -0,0 +1,6 @@
+#!/bin/sh
+# Respawned by init. A store that must not run parks here instead of exiting,
+# so init does not respawn it in a tight loop.
+. /usr/local/lib/sqs-store/sqs-store-lib.sh
+sqs_run
+exec sleep 2147483647
@@ -0,0 +1,5 @@
+#!/bin/sh
+# The only shutdown action: stop the store and unmount its disk, then run the
+# generic read-only remount only if the store is gone.
+. /usr/local/lib/sqs-store/sqs-store-lib.sh
+sqs_shutdown
@@ -0,0 +1,8 @@
+# The power button loads here, not in /init: busybox init has its SIGUSR2
+# (poweroff) handler installed before sysinit runs, so a press can never reach
+# the pre-exec shell PID 1. A press before this line is lost; the host re-sends.
+::sysinit:/sbin/modprobe tiny_power_button power_signal=12
+::sysinit:/usr/local/lib/sqs-worker/sqs-worker-prepare
+::respawn:/usr/local/lib/sqs-worker/sqs-worker-ready
+::ctrlaltdel:/sbin/reboot
+::shutdown:/usr/local/lib/sqs-worker/sqs-worker-stop
@@ -0,0 +1,115 @@
+# sqs-worker-lib.sh — shared by the SQS worker guest's readiness and stop
+# steps. Sourced, never executed. Every path and command is overridable so the
+# host test can drive the same code against test-owned processes and stub tools.
+#
+# store-ready exists only while the running watcher's last check passed. A
+# check proves the store accepted the worker key, not the store server's
+# identity, and a store failure is noticed up to one interval plus call time late.
+
+SQSW_RUN=${SQSW_RUN:-/run/sqs-worker}
+SQSW_DATA=${SQSW_DATA:-/data}
+SQSW_CONSOLE=${SQSW_CONSOLE:-/dev/console}
+SQSW_MOUNTS=${SQSW_MOUNTS:-/proc/mounts}
+SQSW_BIN=${SQSW_BIN:-/usr/local/bin/sqs-worker}
+SQSW_STOP_BUDGET_TENTHS=${SQSW_STOP_BUDGET_TENTHS:-100}
+SQSW_RESPAWN_GAP=${SQSW_RESPAWN_GAP:-5}
+SQSW_PARK_INTERVAL=${SQSW_PARK_INTERVAL:-30}
+UMOUNT=${UMOUNT:-umount}
+
+sqsw_log() { echo "[sqs-worker] $*" >> "$SQSW_CONSOLE"; }
+
+# sqsw_alive <pid>: running or stopped, not exited (a zombie counts as exited).
+sqsw_alive() { [ -r "/proc/$1/status" ] && ! grep -q '^State:[[:space:]]*Z' "/proc/$1/status" 2>/dev/null; }
+
+# sqsw_starttime <pid>: kernel start time (stat field 22). It survives exec, and
+# pairing it with the pid reduces PID-reuse risk; it is not an absolute guarantee.
+sqsw_starttime() { sed 's/.*) //' "/proc/$1/stat" 2>/dev/null | cut -d' ' -f20; }
+
+# sqsw_same <pid> <starttime>: the recorded process is still alive.
+sqsw_same() { [ -n "$2" ] && sqsw_alive "$1" && [ "$(sqsw_starttime "$1")" = "$2" ]; }
+
+# sqsw_run: the respawned readiness body. A predecessor that was killed or
+# crashed could not remove its marker, so every start removes it first. Its
+# identity is recorded before the stopping check, as the store's run step does.
+sqsw_run() {
+    mkdir -p "$SQSW_RUN"
+    echo "$$ $(sqsw_starttime $$)" > "$SQSW_RUN/ready.pid"
+    if [ -e "$SQSW_RUN/store-ready" ]; then
+        rm -f "$SQSW_RUN/store-ready"
+        sqsw_log "ready: removed a store-ready marker left by an earlier watcher"
+    fi
+    # Returning here would make init respawn and log at once, forever. Park
+    # instead, logging once: shutdown stops this recorded process, and a key
+    # that appears later lets the watcher start.
+    if [ -e "$SQSW_RUN/stopping" ] || [ ! -s "$SQSW_RUN/public-key" ]; then
+        if [ -e "$SQSW_RUN/stopping" ]; then
+            sqsw_log "ready: shutdown in progress, not starting"
+        else
+            sqsw_log "ready: no worker key; readiness not started, waiting"
+        fi
+        while [ -e "$SQSW_RUN/stopping" ] || [ ! -s "$SQSW_RUN/public-key" ]; do
+            sleep "$SQSW_PARK_INTERVAL"
+        done
+    fi
+    # Init respawns at once; a watcher that keeps exiting restarts at most
+    # once per gap instead of in a tight loop.
+    now=$(date +%s)
+    last=$(cat "$SQSW_RUN/ready.last" 2>/dev/null || echo 0)
+    wait=$((last + SQSW_RESPAWN_GAP - now))
+    if [ "$wait" -gt 0 ] && [ "$wait" -le "$SQSW_RESPAWN_GAP" ]; then sleep "$wait"; fi
+    date +%s > "$SQSW_RUN/ready.last"
+    sqsw_log "ready: starting watcher pid=$$"
+    exec "$SQSW_BIN" ready >> "$SQSW_CONSOLE" 2>&1
+}
+
+# sqsw_stop: block any respawn, drop the marker, SIGTERM the recorded watcher,
+# SIGKILL only after the budget, and unmount the worker disk only once the
+# watcher is confirmed gone. A survivor keeps the disk mounted.
+sqsw_stop() {
+    mkdir -p "$SQSW_RUN"
+    : > "$SQSW_RUN/stopping"
+    rm -f "$SQSW_RUN/store-ready"
+    rec=$(cat "$SQSW_RUN/ready.pid" 2>/dev/null || true)
+    pid=${rec%% *} st=${rec#* }
+    [ "$st" != "$rec" ] || st=""
+    if [ -n "$pid" ] && sqsw_same "$pid" "$st"; then
+        kill -TERM "$pid" 2>/dev/null || true
+        sqsw_log "stop: SIGTERM sent to readiness watcher pid=$pid"
+        i=0
+        while sqsw_same "$pid" "$st" && [ "$i" -lt "$SQSW_STOP_BUDGET_TENTHS" ]; do sleep 0.1; i=$((i + 1)); done
+        if sqsw_same "$pid" "$st"; then
+            sqsw_log "stop: readiness watcher pid=$pid still running after $((SQSW_STOP_BUDGET_TENTHS / 10))s, SIGKILL"
+            kill -KILL "$pid" 2>/dev/null || true
+            j=0
+            while sqsw_same "$pid" "$st" && [ "$j" -lt 30 ]; do sleep 0.1; j=$((j + 1)); done
+        fi
+        if sqsw_same "$pid" "$st"; then
+            rm -f "$SQSW_RUN/store-ready"
+            sqsw_log "stop: readiness watcher pid=$pid STILL ALIVE, leaving $SQSW_DATA mounted"
+            return 1
+        fi
+        sqsw_log "stop: readiness watcher pid=$pid exited"
+    else
+        sqsw_log "stop: no running readiness watcher"
+    fi
+    rm -f "$SQSW_RUN/store-ready" "$SQSW_RUN/ready.pid"
+    sync
+    if grep -q " $SQSW_DATA " "$SQSW_MOUNTS" 2>/dev/null; then
+        rc=0
+        "$UMOUNT" "$SQSW_DATA" || rc=$?
+        sqsw_log "stop: umount $SQSW_DATA rc=$rc"
+    fi
+}
+
+# sqsw_shutdown: the only shutdown action. The generic read-only remount runs
+# only after the watcher stopped; a survivor keeps every mount as it is.
+sqsw_shutdown() {
+    if ! sqsw_stop; then
+        sync
+        sqsw_log "shutdown: readiness watcher survived, skipping umount -a -r"
+        return 1
+    fi
+    rc=0
+    "$UMOUNT" -a -r || rc=$?
+    echo "[init] shutdown: umount -a -r rc=$rc" >> "$SQSW_CONSOLE"
+}
@@ -0,0 +1,8 @@
+#!/bin/sh
+# sysinit: load the disk drivers, then select, check and mount the worker disk
+# and load or create the worker's store key on it. Any refusal is logged; the
+# key is never read or created on a disk that is not proven mounted.
+modprobe virtio_blk 2>/dev/null || true
+modprobe ext4 2>/dev/null || true
+/usr/local/bin/sqs-worker boot >> /dev/console 2>&1 || echo "[sqs-worker] boot refused" >> /dev/console
+exit 0
@@ -0,0 +1,6 @@
+#!/bin/sh
+# Respawned by init. A watcher that must not run parks here instead of exiting,
+# so init does not respawn it in a tight loop.
+. /usr/local/lib/sqs-worker/sqs-worker-lib.sh
+sqsw_run
+exec sleep 2147483647
@@ -0,0 +1,5 @@
+#!/bin/sh
+# The only shutdown action: stop the readiness watcher and unmount the worker
+# disk, then run the generic read-only remount only if the watcher is gone.
+. /usr/local/lib/sqs-worker/sqs-worker-lib.sh
+sqsw_shutdown
@@ -1,10 +1,10 @@
-# spinifex-unit-version: 3
+# spinifex-unit-version: 5
 [Unit]
 Description=Spinifex Daemon (VM Orchestration)
 PartOf=spinifex.target
-After=spinifex-nats.service spinifex-viperblock.service openvswitch-switch.service ovn-controller.service openvswitch-ipsec.service
-# Wants (not Requires): stay bootable when openvswitch-ipsec is absent
-Wants=openvswitch-ipsec.service
+After=spinifex-nats.service spinifex-viperblock.service ovsdb-server.service ovs-vswitchd.service ovn-controller.service openvswitch-ipsec.service spinifex-hostnet.service
+# Wants (not Requires): boot without openvswitch-ipsec, survive OVS restarts via hostnet
+Wants=openvswitch-ipsec.service spinifex-hostnet.service

 [Service]
 EnvironmentFile=-/etc/spinifex/telemetry.env
@@ -14,6 +14,7 @@ Type=simple
 Slice=spinifex-system.slice
 User=spinifex-daemon
 Group=spinifex
+ExecStartPre=/usr/local/bin/spx admin hostnet check
 ExecStart=/usr/local/bin/spx service spinifex start
 Restart=on-failure
 RestartSec=5
@@ -0,0 +1,25 @@
+# spinifex-unit-version: 2
+[Unit]
+Description=Spinifex host network wiring (veth pairs, mgmt address, encap route)
+# PartOf/WantedBy ovs-vswitchd: re-applied whenever OVS starts; daemon and vpcd only Want it
+Requires=ovs-vswitchd.service
+After=ovs-vswitchd.service NetworkManager-wait-online.service
+PartOf=ovs-vswitchd.service
+# Stopping OVS stops ovn-controller through its Requires= chain and starting OVS
+# does not bring it back, so the start that re-applies hostnet restarts it too
+Wants=ovn-controller.service
+Before=spinifex.target
+
+[Service]
+Type=oneshot
+RemainAfterExit=yes
+ExecStart=/usr/local/bin/spx admin hostnet apply
+# Covers the 120s wait behind setup-ovn.sh plus the 60s wait for links
+TimeoutStartSec=240
+TimeoutStopSec=10
+ProtectSystem=full
+ProtectHome=yes
+PrivateTmp=yes
+
+[Install]
+WantedBy=spinifex.target ovs-vswitchd.service
@@ -1,9 +1,9 @@
-# spinifex-unit-version: 1
+# spinifex-unit-version: 3
 [Unit]
 Description=Spinifex VPC Daemon (OVN Networking)
 PartOf=spinifex.target
-After=spinifex-nats.service spinifex-daemon.service systemd-networkd.service openvswitch-switch.service ovn-controller.service
-Wants=spinifex-nats.service ovn-controller.service
+After=spinifex-nats.service spinifex-daemon.service systemd-networkd.service ovsdb-server.service ovs-vswitchd.service ovn-controller.service spinifex-hostnet.service
+Wants=spinifex-nats.service ovn-controller.service spinifex-hostnet.service

 [Service]
 EnvironmentFile=-/etc/spinifex/telemetry.env
@@ -11,6 +11,7 @@ Type=simple
 Slice=spinifex-system.slice
 User=spinifex-vpcd
 Group=spinifex
+ExecStartPre=/usr/local/bin/spx admin hostnet check
 ExecStartPre=/var/lib/spinifex/wait-for-nats.sh
 ExecStart=/usr/local/bin/spx service vpcd start
 Restart=on-failure
@@ -1,9 +1,9 @@
-# spinifex-unit-version: 1
+# spinifex-unit-version: 4
 [Unit]
 Description=Spinifex Infrastructure Platform
 After=network-online.target
 Wants=network-online.target
-Wants=spinifex-nats.service spinifex-predastore.service spinifex-northstar.service spinifex-viperblock.service spinifex-daemon.service spinifex-awsgw.service spinifex-vpcd.service spinifex-ui.service spinifex-qmp-collector.service spinifex-shutdown.service spinifex-nats-watchdog.timer openvswitch-switch.service ovn-controller.service
+Wants=spinifex-nats.service spinifex-predastore.service spinifex-northstar.service spinifex-viperblock.service spinifex-daemon.service spinifex-awsgw.service spinifex-vpcd.service spinifex-ui.service spinifex-qmp-collector.service spinifex-shutdown.service spinifex-nats-watchdog.timer ovsdb-server.service ovs-vswitchd.service ovn-controller.service

 [Install]
 WantedBy=multi-user.target
@@ -4,11 +4,20 @@ import (
 	"flag"
 	"fmt"
 	"os"
+	"slices"

 	"github.com/mulgadc/predastore/s3api"
 	"github.com/mulgadc/spinifex/internal/awsmodel"
 	acmawsapi "github.com/mulgadc/spinifex/spinifex/domains/acm/awsapi"
+	ec2awsapi "github.com/mulgadc/spinifex/spinifex/domains/ec2/awsapi"
 	awsapi "github.com/mulgadc/spinifex/spinifex/domains/ecr/awsapi"
+	ecsawsapi "github.com/mulgadc/spinifex/spinifex/domains/ecs/awsapi"
+	eksawsapi "github.com/mulgadc/spinifex/spinifex/domains/eks/awsapi"
+	elbv2awsapi "github.com/mulgadc/spinifex/spinifex/domains/elbv2/awsapi"
+	iamawsapi "github.com/mulgadc/spinifex/spinifex/domains/iam/awsapi"
+	rdsawsapi "github.com/mulgadc/spinifex/spinifex/domains/rds/awsapi"
+	route53awsapi "github.com/mulgadc/spinifex/spinifex/domains/route53/awsapi"
+	stsawsapi "github.com/mulgadc/spinifex/spinifex/domains/sts/awsapi"
 	"github.com/mulgadc/spinifex/spinifex/gateway"
 	"github.com/mulgadc/spinifex/spinifex/ingress/aws/dispatch"

@@ -49,9 +58,24 @@ func compareAll() ([]awsmodel.OperationCoverage, error) {
 	// Registered services contribute their declared inventory, which needs no
 	// live dependencies; nothing here serves a request.
 	inventories := gateway.AWSOperationInventory(map[string]dispatch.Inventory{
-		awsapi.ServiceName:    awsapi.OperationInventory(),
-		acmawsapi.ServiceName: acmawsapi.OperationInventory(),
+		awsapi.ServiceName:        awsapi.OperationInventory(),
+		acmawsapi.ServiceName:     acmawsapi.OperationInventory(),
+		route53awsapi.ServiceName: route53awsapi.OperationInventory(),
+		elbv2awsapi.InventoryName: elbv2awsapi.OperationInventory(),
+		rdsawsapi.ServiceName:     rdsawsapi.OperationInventory(),
+		ec2awsapi.ServiceName:     ec2awsapi.OperationInventory(),
+		ecsawsapi.ServiceName:     ecsawsapi.OperationInventory(),
+		eksawsapi.ServiceName:     eksawsapi.OperationInventory(),
+		iamawsapi.ServiceName:     iamawsapi.OperationInventory(),
+		stsawsapi.ServiceName:     stsawsapi.OperationInventory(),
 	})
+	// The loop below walks the embedded models, so a served service with no
+	// model would otherwise drop out of the report without a word.
+	for name := range inventories {
+		if !slices.Contains(awsmodel.Services(), awsmodel.Service(name)) {
+			return nil, fmt.Errorf("no embedded AWS model for served service %s", name)
+		}
+	}
 	coverages := make([]awsmodel.OperationCoverage, 0, len(awsmodel.Services()))
 	for _, service := range awsmodel.Services() {
 		inventory, ok := inventories[string(service)]
@@ -29,6 +29,7 @@ func loadConfig() (tokenwebhook.Config, error) {
 		CertPath:    envOr("EKS_WEBHOOK_CERT", "/etc/spinifex-eks/token-webhook.crt"),
 		KeyPath:     envOr("EKS_WEBHOOK_KEY", "/etc/spinifex-eks/token-webhook.key"),
 		Kubeconfig:  envOr("EKS_WEBHOOK_KUBECONFIG", "/etc/spinifex-eks/token-webhook.kubeconfig"),
+		ReadyAddr:   os.Getenv("EKS_WEBHOOK_READY_ADDR"),
 	}
 	// Static SigV4 creds are optional: when absent, eksgw.New signs with the
 	// AWS SDK chain (IMDS instance-role creds), the same path the CP VM's
@@ -0,0 +1,120 @@
+// sqs-worker runs inside the shared SQS worker microVM.
+//
+// Usage:
+//
+//	sqs-worker boot
+//	sqs-worker ready
+//
+// boot selects and mounts the worker's data disk named by fw_cfg, asks the
+// gateway for its current first-store decision with its instance-role
+// credentials, then loads or creates the store key on that disk and
+// publishes only the public key. ready then keeps checking the bound store's
+// JetStream with that key until it is stopped, keeping a readiness marker only
+// while the check passes; it proves the store accepted the worker key, not
+// the store server's identity.
+package main
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"os"
+	"os/signal"
+	"syscall"
+	"time"
+
+	"github.com/mulgadc/spinifex/internal/gwsign"
+	"github.com/mulgadc/spinifex/internal/sqsworkergw"
+	"github.com/mulgadc/spinifex/spinifex/agents/sqs/workerboot"
+	"github.com/mulgadc/spinifex/spinifex/agents/sqs/workerkey"
+)
+
+// The cluster CA the daemon supplies; the only root the client trusts.
+const caPath = "/sys/firmware/qemu_fw_cfg/by_name/opt/spinifex/ca-cert/raw"
+
+// newClaimer signs with the IMDS instance-role credentials only; there is
+// no static-key fallback.
+func newClaimer() workerboot.NewClaimer {
+	return func(region, gatewayURL string) (workerkey.Claimer, error) {
+		return newClient(region, gatewayURL)
+	}
+}
+
+func newClient(region, gatewayURL string) (*sqsworkergw.Client, error) {
+	ca, err := os.ReadFile(caPath)
+	if err != nil {
+		return nil, fmt.Errorf("read cluster CA: %w", err)
+	}
+	return sqsworkergw.New(gatewayURL, region, ca, gwsign.NewInstanceRole())
+}
+
+type describer struct{ c *sqsworkergw.Client }
+
+func (d describer) DescribeStore(ctx context.Context) (workerboot.Store, error) {
+	st, err := d.c.DescribeStore(ctx)
+	return workerboot.Store(st), err
+}
+
+func main() {
+	if len(os.Args) != 2 || (os.Args[1] != "boot" && os.Args[1] != "ready") {
+		fmt.Fprintln(os.Stderr, "usage: sqs-worker boot|ready")
+		os.Exit(2)
+	}
+	logf := func(msg string) { fmt.Fprintf(os.Stderr, "[sqs-worker] %s\n", msg) }
+	run := boot
+	if os.Args[1] == "ready" {
+		run = ready
+	}
+	if err := run(logf); err != nil {
+		logf("FATAL: " + err.Error())
+		os.Exit(1)
+	}
+}
+
+func config(logf func(string)) workerboot.Config {
+	return workerboot.Config{
+		FwCfgDir:      "/sys/firmware/qemu_fw_cfg/by_name/opt/spinifex/sqs-worker",
+		SysBlock:      "/sys/block",
+		DevDir:        "/dev",
+		MountInfo:     "/proc/self/mountinfo",
+		DataDir:       "/data",
+		RunDir:        "/run/sqs-worker",
+		Disk:          workerboot.SystemDisk{E2fsck: "/sbin/e2fsck"},
+		IsBlockDevice: workerboot.IsBlockDevice,
+		NewClaimer:    newClaimer(),
+		Attempts:      60,
+		Wait:          5 * time.Second,
+		Log:           logf,
+	}
+}
+
+func boot(logf func(string)) error {
+	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Minute)
+	defer cancel()
+	_, err := workerboot.Boot(ctx, config(logf))
+	return err
+}
+
+// ready runs until init stops it; the marker is removed on the way out.
+func ready(logf func(string)) error {
+	ctx, cancel := signal.NotifyContext(context.Background(), syscall.SIGTERM, syscall.SIGINT)
+	defer cancel()
+	err := workerboot.Watch(ctx, workerboot.ReadyConfig{
+		Config: config(logf),
+		NewDescriber: func(region, gatewayURL string) (workerboot.StoreDescriber, error) {
+			c, err := newClient(region, gatewayURL)
+			if err != nil {
+				return nil, err
+			}
+			return describer{c}, nil
+		},
+		Checker:     workerboot.NATSChecker{},
+		CallTimeout: 30 * time.Second,
+		Refresh:     30 * time.Second,
+		Refused:     5 * time.Minute,
+	})
+	if errors.Is(err, context.Canceled) {
+		return nil
+	}
+	return err
+}
@@ -0,0 +1,29 @@
+package clusterv1
+
+import "time"
+
+// ClusterStateBucket is the KV bucket holding each daemon's heartbeat alongside
+// other cluster-wide state.
+const ClusterStateBucket = "spinifex-cluster-state"
+
+// HeartbeatKey is the ClusterStateBucket key a node writes its heartbeat under.
+func HeartbeatKey(node string) string {
+	return "heartbeat." + node
+}
+
+// HeartbeatTimestamp is the read side of a heartbeat for node liveness. It
+// decodes only the timestamp, so a reader does not depend on the capacity
+// figures the daemon writes beside it.
+type HeartbeatTimestamp struct {
+	Timestamp string `json:"timestamp"`
+}
+
+// Time returns when the heartbeat was written. A timestamp that is missing or
+// unparseable vouches for nothing, so ok is false.
+func (h HeartbeatTimestamp) Time() (written time.Time, ok bool) {
+	parsed, err := time.Parse(time.RFC3339, h.Timestamp)
+	if err != nil {
+		return time.Time{}, false
+	}
+	return parsed.UTC(), true
+}
@@ -0,0 +1,35 @@
+package clusterv1_test
+
+import (
+	"encoding/json"
+	"testing"
+	"time"
+
+	clusterv1 "github.com/mulgadc/spinifex/contracts/cluster/v1"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// The bucket and key are the deployed layout: a rolling upgrade has old daemons
+// writing and new readers reading, so these strings cannot change.
+func TestHeartbeatKey_IsTheDeployedLayout(t *testing.T) {
+	assert.Equal(t, "spinifex-cluster-state", clusterv1.ClusterStateBucket)
+	assert.Equal(t, "heartbeat.node-a", clusterv1.HeartbeatKey("node-a"))
+}
+
+func TestHeartbeatTimestamp_ReadsAFullDaemonHeartbeat(t *testing.T) {
+	full := `{"node":"node-a","epoch":3,"timestamp":"2026-10-30T01:02:03Z","vm_count":4,"allocated_vcpu":8}`
+	var h clusterv1.HeartbeatTimestamp
+	require.NoError(t, json.Unmarshal([]byte(full), &h))
+
+	written, ok := h.Time()
+	require.True(t, ok)
+	assert.True(t, written.Equal(time.Date(2026, 10, 30, 1, 2, 3, 0, time.UTC)))
+}
+
+func TestHeartbeatTimestamp_UndatedBeatVouchesForNothing(t *testing.T) {
+	for _, ts := range []string{"", "yesterday"} {
+		_, ok := clusterv1.HeartbeatTimestamp{Timestamp: ts}.Time()
+		assert.False(t, ok, "timestamp %q", ts)
+	}
+}
@@ -0,0 +1,7 @@
+package ec2v1
+
+// VolumeHeldElsewhereText is the text of a start refused because another node
+// holds one of the instance's volumes. It crosses NATS as message text, so
+// readers match on it; the provider that refuses and every reader share this
+// string so they cannot drift.
+const VolumeHeldElsewhereText = "volume is leased by another owner"
@@ -8,3 +8,8 @@ administration. It is separate from AWS tenant-facing service contracts.
 roles and Reed-Solomon configuration. It is an operational status contract,
 not a generic provider abstraction; do not change its subject or JSON shape
 incompatibly. Add a new version when that is required.
+
+`admin.go` holds the `/admin/<Method>` request and response bodies the private admin surface shares with the CLI.
+`AdminMethodNames` is the list of callable methods, and the gateway router builds its accept set from it.
+`AccountQuotaRequest`, `AccountQuotaResponse` and `DeleteAccountResponse` stay in the gateway until the `/admin` handlers move, because they embed types the contracts layer may not import.
+Do not change a field name, JSON tag or omitempty incompatibly; add a new version when that is required.
@@ -0,0 +1,76 @@
+package operatorv1
+
+import "slices"
+
+// adminMethodNames is the callable /admin/<Method> list. The gateway router
+// builds its accept set from it, so a name cannot be served without being listed.
+var adminMethodNames = []string{
+	"CreateAccount",
+	"DeleteAccount",
+	"DescribeAccountDeletion",
+	"GetAccountQuota",
+	"ListAccounts",
+	"PutAccountQuota",
+}
+
+// AdminMethodNames returns the callable /admin/<Method> names in sorted order.
+// Tooling that mints a credential grants these by name, so the list has to be
+// the one the router serves rather than a copy that can drift out of step.
+func AdminMethodNames() []string {
+	return slices.Sorted(slices.Values(adminMethodNames))
+}
+
+// Deletion job states.
+const (
+	DeletionStateRunning   = "RUNNING"
+	DeletionStateCompleted = "COMPLETED"
+	DeletionStateFailed    = "FAILED"
+)
+
+// CreateAccountRequest is the /admin/CreateAccount request body.
+type CreateAccountRequest struct {
+	Name        string `json:"name"`
+	ClientToken string `json:"clientToken"`
+	Source      string `json:"source,omitempty"`
+}
+
+// CreateAccountResponse is the /admin/CreateAccount success body. It is stored
+// verbatim (with the secret encrypted) so a replay returns an identical answer.
+type CreateAccountResponse struct {
+	AccountID       string `json:"accountId"`
+	AccountName     string `json:"accountName"`
+	AdminUser       string `json:"adminUser"`
+	AccessKeyID     string `json:"accessKeyId"`
+	SecretAccessKey string `json:"secretAccessKey"`
+	DefaultVpcID    string `json:"defaultVpcId"`
+}
+
+// DeleteAccountRequest is the /admin/DeleteAccount request body.
+type DeleteAccountRequest struct {
+	AccountID   string `json:"accountId"`
+	AccountName string `json:"accountName"`
+	ClientToken string `json:"clientToken"`
+	Force       bool   `json:"force,omitempty"`
+	DryRun      bool   `json:"dryRun,omitempty"`
+}
+
+// DescribeAccountDeletionRequest is the /admin/DescribeAccountDeletion body.
+type DescribeAccountDeletionRequest struct {
+	AccountID string `json:"accountId"`
+}
+
+// AccountSummary is one row of /admin/ListAccounts. It carries no credential
+// material: the caller can already create and delete accounts, but that is no
+// reason for a listing to hand out keys.
+type AccountSummary struct {
+	AccountID   string `json:"accountId"`
+	AccountName string `json:"accountName"`
+	Status      string `json:"status"`
+	CreatedAt   string `json:"createdAt,omitempty"`
+}
+
+// ListAccountsResponse is the /admin/ListAccounts success body.
+type ListAccountsResponse struct {
+	Accounts []AccountSummary `json:"accounts"`
+	Count    int              `json:"count"`
+}
@@ -0,0 +1,65 @@
+package operatorv1
+
+import (
+	"encoding/json"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+)
+
+func goldenJSON(t *testing.T, v any, want string) {
+	t.Helper()
+	data, err := json.Marshal(v)
+	require.NoError(t, err)
+	require.JSONEq(t, want, string(data))
+}
+
+func TestAdminWire_CreateAccountRequest(t *testing.T) {
+	goldenJSON(t, CreateAccountRequest{Name: "acme", ClientToken: "tok-1", Source: "cli"},
+		`{"name":"acme","clientToken":"tok-1","source":"cli"}`)
+	goldenJSON(t, CreateAccountRequest{Name: "acme", ClientToken: "tok-1"},
+		`{"name":"acme","clientToken":"tok-1"}`)
+}
+
+func TestAdminWire_CreateAccountResponse(t *testing.T) {
+	goldenJSON(t, CreateAccountResponse{
+		AccountID: "000000000042", AccountName: "acme", AdminUser: "admin",
+		AccessKeyID: "AKIAEXAMPLE", SecretAccessKey: "secret", DefaultVpcID: "vpc-1",
+	}, `{"accountId":"000000000042","accountName":"acme","adminUser":"admin","accessKeyId":"AKIAEXAMPLE","secretAccessKey":"secret","defaultVpcId":"vpc-1"}`)
+}
+
+func TestAdminWire_DeleteAccountRequest(t *testing.T) {
+	goldenJSON(t, DeleteAccountRequest{AccountID: "000000000042", AccountName: "acme", ClientToken: "tok-2", Force: true, DryRun: true},
+		`{"accountId":"000000000042","accountName":"acme","clientToken":"tok-2","force":true,"dryRun":true}`)
+	goldenJSON(t, DeleteAccountRequest{AccountID: "000000000042", AccountName: "acme", ClientToken: "tok-2"},
+		`{"accountId":"000000000042","accountName":"acme","clientToken":"tok-2"}`)
+}
+
+func TestAdminWire_DescribeAccountDeletionRequest(t *testing.T) {
+	goldenJSON(t, DescribeAccountDeletionRequest{AccountID: "000000000042"}, `{"accountId":"000000000042"}`)
+}
+
+func TestAdminWire_DeletionStates(t *testing.T) {
+	require.Equal(t, "RUNNING", DeletionStateRunning)
+	require.Equal(t, "COMPLETED", DeletionStateCompleted)
+	require.Equal(t, "FAILED", DeletionStateFailed)
+}
+
+func TestAdminWire_ListAccountsResponse(t *testing.T) {
+	goldenJSON(t, ListAccountsResponse{
+		Accounts: []AccountSummary{
+			{AccountID: "000000000042", AccountName: "acme", Status: "ACTIVE", CreatedAt: "2026-01-02T03:04:05Z"},
+			{AccountID: "000000000043", AccountName: "bolt", Status: "TERMINATING"},
+		},
+		Count: 2,
+	}, `{"accounts":[
+		{"accountId":"000000000042","accountName":"acme","status":"ACTIVE","createdAt":"2026-01-02T03:04:05Z"},
+		{"accountId":"000000000043","accountName":"bolt","status":"TERMINATING"}],"count":2}`)
+}
+
+func TestAdminWire_AdminMethodNames(t *testing.T) {
+	require.Equal(t, []string{
+		"CreateAccount", "DeleteAccount", "DescribeAccountDeletion",
+		"GetAccountQuota", "ListAccounts", "PutAccountQuota",
+	}, AdminMethodNames())
+}
@@ -0,0 +1,71 @@
+package telemetryv1
+
+import (
+	"regexp"
+)
+
+// LiveSamplePeriodMillis is how often a watched instance is sampled for the
+// console's live view.
+const LiveSamplePeriodMillis = 500
+
+// LiveWatchLeaseMillis is how long one watch keeps the live sampler running.
+// Watchers renew well inside it, so a viewer that goes away costs at most this.
+const LiveWatchLeaseMillis = 30_000
+
+// LiveRingSamples bounds the in-memory history a live sampler keeps for a
+// reconnecting viewer: five minutes at LiveSamplePeriodMillis.
+const LiveRingSamples = 5 * 60 * 1000 / LiveSamplePeriodMillis
+
+// Live subjects sit outside metrics.> so nothing that ingests the periodic
+// CloudWatch batches ever reads a sub-second rate as a period. The account is part
+// of the subject, so only the poller for that account's instance answers.
+const (
+	liveWatchSubjectPrefix  = "telemetry.watch.ec2."
+	liveSampleSubjectPrefix = "telemetry.live.ec2."
+)
+
+var (
+	liveAccountPattern  = regexp.MustCompile(`^[0-9]{12}$`)
+	liveInstancePattern = regexp.MustCompile(`^i-[0-9a-f]{8,17}$`)
+)
+
+// ValidLiveTarget reports whether accountID and instanceID are safe to place
+// in a live subject: no NATS token separators or wildcards can get through.
+func ValidLiveTarget(accountID, instanceID string) bool {
+	return liveAccountPattern.MatchString(accountID) && liveInstancePattern.MatchString(instanceID)
+}
+
+// LiveWatchSubject is where a viewer asks the owning node to keep sampling. A
+// request on it is answered with the ring since LiveWatch.SinceMillis.
+func LiveWatchSubject(accountID, instanceID string) string {
+	return liveWatchSubjectPrefix + accountID + "." + instanceID
+}
+
+// LiveSampleSubject carries one LiveSample per sampling period while watched.
+func LiveSampleSubject(accountID, instanceID string) string {
+	return liveSampleSubjectPrefix + accountID + "." + instanceID
+}
+
+// LiveWatch is the body of a watch. SinceMillis is the newest sample the
+// viewer already has; zero asks for the whole ring.
+type LiveWatch struct {
+	SinceMillis int64 `json:"since_ms,omitempty"`
+}
+
+// LiveRing answers a watch request with the samples newer than SinceMillis.
+type LiveRing struct {
+	Samples []LiveSample `json:"samples"`
+}
+
+// LiveSample is one live measurement. Every value is a rate over the interval
+// ending at TimeMillis, measured against the previous sample's clock.
+type LiveSample struct {
+	TimeMillis int64   `json:"t"`
+	CPUPercent float64 `json:"cpu"`
+	// NetworkIn and NetworkOut are bytes per second from the guest's side.
+	NetworkIn  float64 `json:"netIn"`
+	NetworkOut float64 `json:"netOut"`
+	// PacketsIn and PacketsOut are packets per second from the guest's side.
+	PacketsIn  float64 `json:"pktIn"`
+	PacketsOut float64 `json:"pktOut"`
+}
@@ -0,0 +1,42 @@
+package telemetryv1
+
+import (
+	"encoding/json"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// The console parses these names, so a rename here breaks its live charts.
+func TestLiveSampleJSONContract(t *testing.T) {
+	data, err := json.Marshal(LiveSample{
+		TimeMillis: 1793450123250, CPUPercent: 12.5,
+		NetworkIn: 48213, NetworkOut: 1290, PacketsIn: 41, PacketsOut: 12,
+	})
+	require.NoError(t, err)
+	require.JSONEq(t, `{"t":1793450123250,"cpu":12.5,"netIn":48213,"netOut":1290,"pktIn":41,"pktOut":12}`, string(data))
+}
+
+func TestValidLiveTarget(t *testing.T) {
+	assert.True(t, ValidLiveTarget("123456789012", "i-0123456789abcdef0"))
+	assert.True(t, ValidLiveTarget("123456789012", "i-0abc1234"))
+
+	for _, tc := range []struct{ account, instance string }{
+		{"", "i-0123456789abcdef0"},
+		{"12345678901", "i-0123456789abcdef0"},
+		{"123456789012", ""},
+		{"123456789012", ">"},
+		{"123456789012", "*"},
+		{"123456789012", "i-0123.abcdef"},
+		{"123456789012", "i-0123456789ABCDEF0"},
+		{"123456789012", "vol-0123456789abcdef0"},
+	} {
+		assert.False(t, ValidLiveTarget(tc.account, tc.instance), "%q %q", tc.account, tc.instance)
+	}
+}
+
+func TestLiveSubjectsAreOutsideMetrics(t *testing.T) {
+	assert.Equal(t, "telemetry.watch.ec2.123456789012.i-0abc1234", LiveWatchSubject("123456789012", "i-0abc1234"))
+	assert.Equal(t, "telemetry.live.ec2.123456789012.i-0abc1234", LiveSampleSubject("123456789012", "i-0abc1234"))
+}
@@ -95,7 +95,7 @@ Operational commands for inspecting cluster state. These fan out NATS requests t

 | Command | Flags | Description |
 |---------|-------|-------------|
-| `spx admin init` | `--nodes`, `--node`, `--bind`, `--advertise` (external IP off-host clients dial; auto-detected from the default route), `--port`, `--region`, `--az`, `--cluster-name`, `--cluster-bind`, `--cluster-routes`, `--predastore-nodes`, `--predastore-compaction-interval` (seconds; test clusters set a short one), `--services`, `--formation-timeout`, `--token-ttl`, `--force`, `--discard-jetstream` (with `--nodes` >= 2, remove this node's existing JetStream streams once every node has joined; without it a multi-node init over a store that holds streams is refused, and it never proceeds while NATS is running), `--skip-host-dns`, `--email` (operator address for update and security notices), `--no-telemetry`, `--gpu-passthrough`, `--ipsec` (default true; encrypt intra-AZ Geneve via OVN native IPsec cluster-wide — disable only for a trusted single-rack lab), `--host` (leader to join; falls back to multicast discovery), `--external-mode` (`pool` \| `nat` routed NAT for non-bridgeable uplinks, single-node; pair with `setup-ovn.sh --nat-uplink`; note: system instances (ECS/EKS/load-balancer agents) are unsupported in nat v1 — they require `pool` mode), `--external-iface` (WAN NIC for br-external, auto-detected), `--external-source` (`dhcp` \| `static`), `--external-pool` (`start-end`), `--external-prefix-len` (default 24), `--external-gateway` (auto-detected), `--external-dns` (comma-separated IPv4 addresses, at most 3; the DNS servers written to the external pool's `dns_servers` and handed to guests, verbatim; when omitted they are detected from the host resolver at init time, so set it to give every node of a cluster the same value), `--external-bind-bridge` (default `br-wan` when `--external-source=dhcp`), `--gateway-ip` (OVN gateway router's external IP for SNAT; defaults to the pool's range start, required for `nat` mode without DHCP) | Generates root IAM credentials (AKIA-prefixed access key + secret) → creates master.key (AES-256, 32 bytes, 0600) → writes bootstrap.json (consumed on first start) → generates CA + server TLS certificates → generates join token (written to `join-token` file, displayed in join command) → creates NATS config with auth token → writes spinifex.toml, awsgw.toml, predastore.toml → configures AWS CLI `spx` profile → creates directory structure under `~/spinifex/` → points the host resolver at this node's local northstar so the Spinifex zones (ELB/EKS names) resolve from the node itself (opt out with `--skip-host-dns`). `--force` on an already-initialized node is idempotent for crypto: it preserves the existing keys (master.key, predastore/viperblock encryption keys), system + admin credentials, and CA, and refreshes only the config files and the CA-signed server certificate (so a changed bind IP / SANs is picked up without breaking already-joined nodes or CA-baked AMIs). A genuine clean slate requires removing the data dirs first. |
+| `spx admin init` | `--nodes`, `--node`, `--bind`, `--advertise` (external IP off-host clients dial; auto-detected from the default route), `--port`, `--region`, `--az`, `--cluster-name`, `--cluster-bind`, `--cluster-routes`, `--predastore-nodes`, `--predastore-compaction-interval` (seconds; test clusters set a short one), `--services`, `--formation-timeout`, `--token-ttl`, `--force`, `--discard-jetstream` (with `--nodes` >= 2, remove this node's existing JetStream streams once every node has joined; without it a multi-node init over a store that holds streams is refused, and it never proceeds while NATS is running), `--skip-host-dns`, `--email` (operator address for update and security notices), `--no-telemetry`, `--gpu-passthrough`, `--ipsec` (default true; encrypt intra-AZ Geneve via OVN native IPsec cluster-wide — disable only for a trusted single-rack lab; on a host whose IPsec stack is libreswan it defaults to false and `--ipsec=true` is refused), `--host` (leader to join; falls back to multicast discovery), `--external-mode` (`pool` \| `nat` routed NAT for non-bridgeable uplinks, single-node; pair with `setup-ovn.sh --nat-uplink`; note: system instances (ECS/EKS/load-balancer agents) are unsupported in nat v1 — they require `pool` mode), `--external-iface` (WAN NIC for br-external, auto-detected), `--external-source` (`dhcp` \| `static`), `--external-pool` (`start-end`), `--external-prefix-len` (default 24), `--external-gateway` (auto-detected), `--external-dns` (comma-separated IPv4 addresses, at most 3; the DNS servers written to the external pool's `dns_servers` and handed to guests, verbatim; when omitted they are detected from the host resolver at init time, so set it to give every node of a cluster the same value), `--external-bind-bridge` (default `br-wan` when `--external-source=dhcp`), `--gateway-ip` (OVN gateway router's external IP for SNAT; defaults to the pool's range start, required for `nat` mode without DHCP) | Generates root IAM credentials (AKIA-prefixed access key + secret) → creates master.key (AES-256, 32 bytes, 0600) → writes bootstrap.json (consumed on first start) → generates CA + server TLS certificates → generates join token (written to `join-token` file, displayed in join command) → creates NATS config with auth token → writes spinifex.toml, awsgw.toml, predastore.toml → configures AWS CLI `spx` profile → creates directory structure under `~/spinifex/` → points the host resolver at this node's local northstar so the Spinifex zones (ELB/EKS names) resolve from the node itself (opt out with `--skip-host-dns`). `--force` on an already-initialized node is idempotent for crypto: it preserves the existing keys (master.key, predastore/viperblock encryption keys), system + admin credentials, and CA, and refreshes only the config files and the CA-signed server certificate (so a changed bind IP / SANs is picked up without breaking already-joined nodes or CA-baked AMIs). A genuine clean slate requires removing the data dirs first. |
 | `spx admin join` | `--host` (required), `--node` (required), `--token` (required), `--bind`, `--advertise` (defaults to `--bind`, or the auto-detected WAN IP when `--bind` is 0.0.0.0), `--port`, `--region`, `--az`, `--cluster-bind`, `--cluster-routes`, `--data-dir`, `--services`, `--skip-host-dns`, `--email`, `--no-telemetry`, `--predastore-compaction-interval`, `--join-timeout` (how long to keep retrying while the formation server is unreachable, default 20m), `--force` (join even though this node is already initialized or holds JetStream streams, **discarding its own CA, master key and JetStream store**; the store is removed only after formation succeeds, and the join is refused while NATS is running even with `--force`) | Connects to leader node with join token (Authorization: Bearer header) → retrieves cluster configuration → configures local node to join cluster and participate in distributed operations → points the host resolver at this node's local northstar so the Spinifex zones resolve from the node itself (opt out with `--skip-host-dns`) |

 ### Version
@@ -127,7 +127,7 @@ Operational commands for inspecting cluster state. These fan out NATS requests t

 | Command | Flags | Description |
 |---------|-------|-------------|
-| `spx admin preflight` | `--quiet` (print only non-OK rows) | Reports whether this host has the privileged helper scripts and sudoers grants the installed build depends on. `spx version` knows the build but not the host: `setup.sh` and `setup-ovn.sh` install those assets, and a binary-swap deploy (`make deploy`, or a manual copy) skips them. It also reports any running service still executing a replaced binary — overwriting `/usr/local/bin/spx` while `spinifex.target` is up does not move services onto it, so each keeps executing the replaced inode until it restarts and a node can serve a mixture of two builds indefinitely. Checks compare content, not existence, so a present-but-stale asset is caught. Exits non-zero if any asset is Missing, Stale or Ungranted. |
+| `spx admin preflight` | `--quiet` (print only non-OK rows) | Reports whether this host has the privileged helper scripts and sudoers grants the installed build depends on. `spx version` knows the build but not the host: `setup.sh` and `setup-ovn.sh` install those assets, and a binary-swap deploy (`make deploy`, or a manual copy) skips them. It also reports any running service still executing a replaced binary — overwriting `/usr/local/bin/spx` while `spinifex.target` is up does not move services onto it, so each keeps executing the replaced inode until it restarts and a node can serve a mixture of two builds indefinitely. Checks compare content, not existence, so a present-but-stale asset is caught. On x86_64 it also checks that the QEMU binary the daemon will run lists the `microvm` machine type and that the microvm firmware `/usr/share/spinifex/microvm/qboot.rom` is installed, since load balancers boot on both. It runs the host network check `spinifex-daemon` and `spinifex-vpcd` run before they start, and fails while firewalld or ufw is active, since the Spinifex nftables policy is the only supported node firewall. It reports a Warning, without failing, when systemd-resolved is active but `/etc/resolv.conf` is not its stub, since lookups then bypass the route that sends Spinifex names to northstar. Exits non-zero if any asset is Missing, Stale, Ungranted, Unsupported or Failed. |
 | `spx admin node js-probe` | `--timeout` (default: 10s) | Round-trips a canary key through the cluster-state KV bucket and exits non-zero if the write or the read-back fails. JetStream can latch a write error per message block and keep serving reads — nats-server's filestore records the failure on the block and returns it on every later write, and nothing clears it short of a restart. The HTTP health endpoint reports the server-wide JetStream state, which that latch never touches, so only an actual write observes it. |
 | `spx admin banner` | `--boot-check` | Writes the node information banner to `/etc/issue` (shown before login on the physical/serial console) and `/etc/motd` (shown after SSH login). With `--boot-check` it also detects a management IP that has changed since last boot and updates `/etc/spinifex/node.conf` accordingly. |

@@ -136,7 +136,7 @@ Operational commands for inspecting cluster state. These fan out NATS requests t
 | Command | Flags | Description |
 |---------|-------|-------------|
 | `spx admin volumes orphans` | — | Lists volumes the storage provider holds that the control plane has no record of. Their blocks consume space but they have no API handle, so they can be neither described nor deleted through the EC2 API. **Reports only, never deletes**: the evidence that an orphan is still wanted is the record that went missing, so removal is an operator decision taken per volume. |
-| `spx admin volumes unsealed` | — | Lists volumes a node has open, or had open, without a confirmed seal to the backend — the node named holds writes predastore may not have. Starting such a volume elsewhere is allowed, and is preferred over leaving the instance unable to run, but it opens from the last checkpoint that did reach the backend. This is the list of volumes where that would cost something. |
+| `spx admin volumes unsealed` | — | Lists volumes a node has open, or had open, without a confirmed seal to the backend — the node named holds writes predastore may not have. Starting such a volume elsewhere is refused while that node is alive (its daemon heartbeat is under 60 seconds old, or cannot be read) and is otherwise allowed, in which case it opens from the last checkpoint that did reach the backend. This is the list of volumes where that would cost something. |
 | `spx admin storage status` | `--endpoint` (default: this node's AWS gateway), `--region` (default: this node's region), `--ca-bundle` (default: this node's CA) | Sends the signed `GetStorageStatus` action to the AWS gateway with credentials from the standard AWS chain (e.g. `AWS_PROFILE=spinifex`) and prints what the console's Service Metrics page shows: predastore encoding, meta and blob node counts, cluster health, a meta node table (health, raft state, leader, term, commit and applied index) and the blob nodes. Cluster health is `Degraded` when any meta node is unreachable or the nodes reporting raft state do not agree on one leader. A meta node read only through its host's `/readyz` shows healthy with blank raft fields. |

 ### Upgrade Management
@@ -181,7 +181,7 @@ Operational commands for inspecting cluster state. These fan out NATS requests t
 | `spx admin gpu status` | `--node` (default: local node) | Queries `spinifex.node.status` fan-out → finds the target node response → prints Node, GPU hardware (model list or "none detected"), IOMMU state, vfio-pci state, passthrough enabled/disabled, and GPU pool allocation (`allocated/total`). Also lists GPU-capable instance types when passthrough is active. |
 | `spx admin gpu enable` | — | Checks current passthrough state via NATS → errors if already enabled or prerequisites not met (directs to `setup`) → writes `gpu_passthrough = true` to `spinifex.toml` via `admin.SetGPUPassthrough` → sends SIGHUP to `spinifex-daemon` via `systemctl kill -s HUP` → polls node status for up to 30 s until daemon confirms new state → prints final `gpu status` output. Must be run directly on the target host. |
 | `spx admin gpu disable` | — | Checks current passthrough state via NATS → errors if already disabled or if `AllocGPUs > 0` (must terminate GPU instances first) → writes `gpu_passthrough = false` → sends SIGHUP to `spinifex-daemon` → polls for up to 30 s → prints final `gpu status` output. Must be run directly on the target host. |
-| `spx admin gpu setup` | — | Idempotent host configuration for GPU passthrough. Steps (skipped if already applied): detect GPUs via `gpu.Discover` → collect PCI IDs for all IOMMU-group siblings → check/enable IOMMU in GRUB (`intel_iommu=on iommu=pt` or `amd_iommu=on iommu=pt`) → write vfio udev rule (`/etc/udev/rules.d/99-spinifex-vfio.rules`) → blacklist nouveau (`/etc/modprobe.d/blacklist-nouveau.conf`) → blacklist amdgpu if AMD GPU present (`/etc/modprobe.d/blacklist-amdgpu.conf`) → write vfio-pci early binding config (`/etc/modprobe.d/vfio-pci.conf`) → add vfio modules to initramfs. If any change requires a reboot: runs `update-initramfs -u` and exits with reboot instructions. After reboot: verifies `vfio_pci` module is loaded → verifies each GPU is bound to `vfio-pci` (binds explicitly via `driver_override` if unbound) → calls `gpu enable` to activate passthrough. |
+| `spx admin gpu setup` | — | Idempotent host configuration for GPU passthrough. Resolves its boot tools first (`update-grub` or `grubby`, `update-initramfs` or `dracut`) and stops before any change if either pair has neither or both. Steps (skipped if already applied): detect GPUs via `gpu.Discover` → collect PCI IDs for all IOMMU-group siblings → if IOMMU is inactive, add `intel_iommu=on iommu=pt` or `amd_iommu=on iommu=pt` to `GRUB_CMDLINE_LINUX_DEFAULT` and run `update-grub`, or run `grubby --update-kernel=ALL --args=...` and check the default entry and `/etc/kernel/cmdline` carry them → write vfio udev rule (`/etc/udev/rules.d/99-spinifex-vfio.rules`) → blacklist nouveau (`/etc/modprobe.d/blacklist-nouveau.conf`) → blacklist amdgpu if AMD GPU present (`/etc/modprobe.d/blacklist-amdgpu.conf`) → write vfio-pci early binding config (`/etc/modprobe.d/vfio-pci.conf`) → add vfio modules to `/etc/initramfs-tools/modules`, or write `/etc/dracut.conf.d/spinifex-vfio.conf` (`force_drivers` for the vfio modules, `install_items` for the `modprobe.d` files). If any change requires a reboot: runs `update-initramfs -u`, or `dracut -f --regenerate-all` and checks every kernel's image with `lsinitrd`, then exits with reboot instructions. After reboot: verifies `vfio_pci` module is loaded → verifies every PCI function in each GPU's slot is bound to `vfio-pci` (binds explicitly via `driver_override` if unbound, fails naming any function on another driver) → calls `gpu enable` to activate passthrough. |
 | `spx admin gpu mig status` | — (must be run on target host) | Runs `gpu.Discover()` locally → for each GPU prints: PCI address, model, MIG capability, MIG mode (enabled/disabled/N/A); for GPUs with MIG enabled lists active MIG slices with GI ID, profile name, and mdev path. |
 | `spx admin gpu mig enable` | `--profile <name>` (required, e.g. `1g.10gb`), `--gpu <pci-addr>` (optional, default: all MIG-capable GPUs) | Checks no GPU instances running (NATS) → discovers MIG-capable GPUs via `gpu.Discover()` (filtered by `--gpu` if set) → enables MIG mode on each target (`gpu.EnableMIGMode`) → lists available profiles (`gpu.ListProfiles`) and validates requested profile name → destroys any existing instances (`gpu.DestroyAllInstances`) → creates new instances filling GPU capacity (`gpu.CreateInstances`) → writes `mig_profile` to `spinifex.toml` via `admin.SetMIGProfile` → sends SIGHUP to `spinifex-daemon`. Must be run directly on the target host. |
 | `spx admin gpu mig disable` | `--gpu <pci-addr>` (optional, default: all MIG-capable GPUs) | Checks no GPU instances running (NATS) → discovers MIG-capable GPUs via `gpu.Discover()` (filtered by `--gpu` if set) → destroys all GPU instances (`gpu.DestroyAllInstances`) → disables MIG mode (`gpu.DisableMIGMode`) → clears `mig_profile` in `spinifex.toml` via `admin.SetMIGProfile` → sends SIGHUP to `spinifex-daemon`. Must be run directly on the target host. |
@@ -192,6 +192,21 @@ Operational commands for inspecting cluster state. These fan out NATS requests t
 |---------|-------|-------------|
 | `spx admin eks restore-snapshot` | `--cluster` (required), `--snapshot` (optional, defaults to the latest snapshot in predastore), `--account` (optional, defaults to the bootstrap account) | Single-CP total-loss DR path (fail-safe): validates the snapshot exists in predastore BEFORE any mutation (a typo'd/missing key hard-fails, never resets into an empty datastore) → launches a fresh control-plane VM as a cluster-init seed (replaying the persisted create-time launch template) → sets a required-snapshot `RecoveryDirective` (`cluster-reset`) so the boot-time recovery agent aborts rather than resets-into-empty if it cannot fetch the snapshot → persists the replacement in cluster meta BEFORE re-pointing the NLB (so an NLB failure is convergeable by the reconciler, returned as a provisional status, not a hard error) → re-points the cluster NLB's apiserver and konnectivity target groups from the old CP's ENI to the new one → fences the old CP with retries, failing loudly if it cannot be confirmed terminated (split-brain guard). Any failure before the meta commit unwinds the fresh CP (terminate + clear directive) so a re-run does not stack a second resetting control plane. The returned status is provisional — success means the sequence completed, not that etcd is restored and serving; verify cluster health. HA clusters (a spread with a potentially surviving quorum) are rejected — recover those via quorum reformation instead. |

+### SQS System Instances
+
+These commands start the first SQS system instances on one node. They do not serve the SQS API, and a bound VM is not a ready service.
+
+Both send a signed `spinifex` admin action (`EnsureSQSWorker`, `EnsureSQSStore`) to the AWS gateway using the standard AWS credential chain, for example `AWS_PROFILE=spinifex`. The caller must be in the super-admin account (`000000000001`) and hold `spinifex:<Action>`. The VMs, volumes, private VPC and ENIs belong to the system account (`000000000000`). `--endpoint`, `--region` and `--ca-bundle` default to this node's gateway, region and `config/ca.pem`; off a node, `--endpoint` and `--region` are required. Each call waits up to 11 minutes.
+
+| Command | Flags | Description |
+|---------|-------|-------------|
+| `spx admin sqs ensure-worker` | `--node` (required), `--instance-type` (required), `--endpoint`, `--region`, `--ca-bundle` | Launches the Region's one SQS worker VM on `--node`, or reports the VM already bound there: node, region, instance, instance type and generation. The worker is pinned to its first node; another node is refused with `InvalidParameterValue`, and there is no takeover. It gets a private system VPC with one /24 subnet and no internet gateway, NAT or Elastic IP, and an encrypted 1 GiB system-account data volume on which the guest generates its own store key. The platform records only the public key. A different `--instance-type` is accepted only while no volume, network interface or launch step has been attempted; once any has, it is `IncorrectState`, whether or not the step completed. |
+| `spx admin sqs ensure-store` | `--node` (required, the worker's node), `--instance-type` (required, more than 512 MiB memory), `--volume-size-gib` (required, 1–16384), `--endpoint`, `--region`, `--ca-bundle` | Launches the Region's SQS store VM on the worker's node with an encrypted system-account data volume, or reports the VM already bound there: node, region, instance, instance type and address. Needs the worker bound with its public key recorded, else `IncorrectState`. The store accepts only that key. A different size, type or key than the first call is `IncorrectState`. |
+
+`ServiceUnavailable` never proves that no SQS resources exist, because an earlier attempt may already have created some. Its message says what it does establish. When no node serves the subject, this request was not delivered. When a volume create was not sent, that particular create did not happen. When the node cannot boot the guest yet or its launcher lacks a capability, this request stopped before allocating or recording anything. When the gateway times out waiting for the node's reply, nothing is known: the launch may still be running, so never assume it created nothing. In every case, call the command again with the same arguments. A retry first checks the durable record of the attempt and continues only when that is safe; it can proceed to the first launch after a create that was proven unsent, and otherwise reports the recorded outcome as bound or `IncorrectState`, or returns `ServiceUnavailable` again. `IncorrectState` from a volume, network interface or launch step means that step was recorded as sent but its outcome was never observed. It is never resent, the message names the volume when its ID is known, and no tool reconciles it yet: recovery is manual.
+
+`bound` says only that the VM was launched for the current record. The worker decides store readiness itself, by checking that the store is bound to its own public key and that a NATS connection with that key succeeds; nothing reports that readiness to the operator. The worker-to-store connection does not authenticate the store server. Real VPC routing, IMDS and gateway TLS from the guest, and worker-to-store traffic on a real interface have not been tested.
+
 ## Private Admin API

 Not an AWS API: a Spinifex-internal surface on the AWS gateway, JSON in and JSON out, one method per path segment at `POST /admin/<Method>`. It exists so a self-service signup form, a provisioning system or a load-test harness can create and remove tenant accounts without an operator shell on a node.
@@ -181,12 +181,17 @@ type Daemon struct {

 Tracks available and allocated CPU/memory to prevent overcommit, and drives the dynamic `ec2.RunInstances.{type}` subscriptions:

+One instance vCPU is one CPU the kernel has online: a thread with SMT on, a core with SMT off or on ARM, as AWS counts a vCPU.
+The OS and Spinifex's own services keep 1 CPU on a host with up to 8, 2 up to 16, 4 up to 63 and 8 from 64, and every other CPU is schedulable one-to-one with no overcommit.
+[ADR-0015](https://github.com/mulgadc/mulga/blob/main/docs/adr/0015-host-cpu-capacity-and-reservation.md) fixes this rule.
+
+
 ```go
 type ResourceManager struct {
    mu            sync.RWMutex
-    hostVCPU      int                                // raw runtime.NumCPU
+    hostVCPU      int                                // online CPUs (ADR-0015:S1)
    hostMemGB     float64                            // raw /proc/meminfo MemTotal
-    reservedVCPU  int                                // held back for spinifex services
+    reservedVCPU  int                                // ADR-0015:S2 tier: 1, 2, 4 or 8
    reservedMem   float64                            // held back for spinifex services
    allocatedVCPU int
    allocatedMem  float64
@@ -229,6 +234,53 @@ type EBSRequest struct {

 Viperblock responds with an NBD URI that QEMU uses to access the block device.

+#### Direct-boot microVM data volumes
+
+A direct-boot microVM keeps the launcher's saved `vm.Config` as a clean base, and `startQEMU` adds data volumes only to a launch-only copy of it (`runtime/compute/vm/lifecycle.go`).
+At each start the daemon takes one snapshot of the instance's already-mounted data requests, probes each NBD or host block export from that snapshot, and builds the command from the same snapshot, so the launched endpoint is the one that was checked and is never written back into the saved base.
+Each data volume becomes a `virtio-blk-device` on the MMIO bus with the same id, node name, I/O thread, serial and `werror=stop,rerror=stop` policy as the PCI device, and takes no PCIe hot-plug port.
+
+Boot and EFI requests are rejected on this path, and a direct-boot instance with no data requests launches its base unchanged.
+Disk-boot (q35) instances keep the existing PCI hot-plug path.
+
+Hot-attaching a volume to a running direct-boot microVM is not supported, because the MMIO bus has no PCIe root port to plug into.
+`AttachVolume` refuses it before mounting the volume or changing any instance or volume state, and the EC2 API returns `UnsupportedOperation`.
+
+This covers only requests that are already mounted when the instance starts; nothing yet supplies data volumes to a system instance, and the path is verified with fake-QEMU unit tests, not with a real EBS volume in a guest.
+
+#### Volume leases
+
+Before opening an engine on a volume, the Viperblock worker takes a lease: one entry per volume in the `VIPERBLOCK_VOLUME_LEASES` JetStream KV bucket, which expires 45 seconds after its last write (`providers/ebs/viperblock/volume_lease.go`).
+A renewal becomes eligible once the last confirmed write is 10 seconds old, and the holder checks for that periodically, so the rewrite lands some time after that point.
+Each rewrite is conditioned on the revision the holder last wrote, and a refused write that finds the entry gone or naming another owner or generation fences the export as a takeover.
+It may keep writing for 40 seconds after the last write the store acknowledged, measured from when that write was sent, because the server starts the entry's expiry when it applies the write and never earlier.
+A slow acknowledgement therefore shortens the remaining validity rather than extending it, and an acquisition acknowledged after its validity has already run out is refused as retryable.
+
+For a holder that cannot renew, a timer is armed for the validity deadline, leaving a nominal 5 seconds before the entry can expire; when it fires, the export's nbdkit is fenced with SIGKILL (`volume_fence.go`).
+The fence runs on its own goroutine, so Go scheduling, that goroutine and process termination all delay the kill, and enough delay can use up the whole margin; a kill held in uninterruptible sleep, which the fence waits up to 30 seconds for, is the extreme case.
+Once a lease is lost or has lapsed it stays lost: a renewal that succeeds afterwards is ignored, and another open on the same node cannot share it while the fence runs.
+
+These rules bound only the holder's own clock.
+The storage backend does not reject a stale writer, so a paused or partitioned host that resumes after its validity can still write; generation enforcement at the write boundary is required by Q-24 in the umbrella repository's `docs/specs/viperblock/rfc3-1_overview.md` and is not implemented.
+An export that survives a daemon restart writes without a renewed lease until recovery adopts it, and a survivor that recovery cannot adopt while a guest still uses it is left running without a lease or a fence.
+
+#### Unsealed writes and start elsewhere (interim)
+
+A volume is unsealed while its node holds writes the backend has not confirmed; the node records this in a dirty marker naming itself as owner (`providers/ebs/viperblock/volume_dirty.go`).
+Opening such a volume on another node loses those writes, so two guards stop it while the previous owner may still hold them.
+This is an interim guard until the instance lifecycle work defines ownership, and it is not a fence.
+
+When `ec2.start` is forwarded to the node that last ran a stopped instance and that node refuses for capacity, the forwarding node normally starts the instance locally.
+If the refusing node still holds unsealed writes for any of the instance's volumes, it sets the `X-Unsealed-Writes` NATS reply header and the forwarder relays `InsufficientInstanceCapacity` instead of starting elsewhere.
+The check asks the owner's viperblockd over the per-node `unsealed` verb with a 5 second bound, and a check that fails or times out counts as holding, so the refusal is kept.
+
+Independently, a node that finds another node's dirty marker on a volume refuses to open it when that node's daemon heartbeat (`heartbeat.<node>` in the cluster-state bucket) is under 60 seconds old.
+It also refuses when the heartbeat cannot be read or no heartbeat source is wired, and the refusal maps to the held-elsewhere error that recovery retries.
+A previous owner that has stopped heartbeating is opened over with a warning, as before.
+The heartbeat age is compared across node clocks, so the 60 second window is also the clock-skew allowance.
+
+EKS control-plane recovery treats both refusals, capacity and held-elsewhere, as retryable: it keeps retrying the stopped member and does not count them against its restart attempts.
+
 ### Predastore (S3)

 Object storage used for:
@@ -256,6 +308,148 @@ file. It does not connect to Predastore to do so, and no Predastore node
 exposes a status endpoint. This is an operator status interface, not a generic
 storage-provider contract.

+### SQS message store (internal)
+
+`domains/sqs/message` defines the store-neutral interface the SQS domain reaches its message store through (ADR-0011:S6), with a JetStream adapter in `domains/sqs/message/jetstream` and a shared conformance suite in `domains/sqs/message/conformance`.
+It is an internal primitive: nothing outside the SQS domain uses it, no daemon or service is wired to it, and it does not itself deliver an SQS API.
+The dedicated JetStream server ADR-0011 describes is not deployed by this code, and Spinifex ships no host service, unit, user or firewall rule for it; the guest that could run it is described under *SQS store system instance (not wired)*.
+
+The interface covers durable and idempotent append, body and metadata reads, conditional state writes and deletes, per-namespace counts, namespace enumeration, an ordered change feed, a bounded paginated scan, namespace deletion and a sweep of tombstoned bodies.
+A namespace carries an immutable message-shard and state-shard allocation chosen when the queue is created, so a later change in shard counts never moves an existing queue.
+
+The adapter provisions a fixed set of shared message streams (`SQS_M_<shard>`, one subject per queue) and shared state buckets (`sqs_state_<shard>`), keeps historical ones, and applies a 14-day-plus-one-hour retention backstop to both.
+It supports one node (R1) and three nodes (R3) only and refuses any other declared cluster size before provisioning.
+
+Reads that must observe a completed write or delete go to the stream leader: `GetState`, the state-key pass behind `Count` and `Namespaces`, and the body pass behind `Namespaces` and `Scan`.
+This avoids the follower-served direct gets and follower-placed consumers that returned purged data in local three-node fault-injection tests; it is not a general linearizability guarantee.
+Each leader read is one request per record, and the body pass transfers each full body internally even though `Scan` returns metadata only, which is a material cost for a large backlog.
+A state-key pass is bounded but is not a snapshot when a key is overwritten during it.
+
+The feed reads through an in-memory, headers-only ordered consumer.
+If that consumer delivers nothing for about two seconds while the stream holds a change after the feed's cursor, the feed replaces it from the cursor, so a stored change does not wait on a stalled consumer; the cause of those stalls is not established.
+`Close` is terminal, waits for consumer opens and deletes in progress and retries failed deletes, but cleanup is best effort for a create whose outcome is unknown: a consumer that lands after its last delete is removed by the server's one-minute inactive threshold.
+
+Coverage is unit and conformance tests on in-process R1 and R3 servers; there is no end-to-end or multi-host validation of this code, and earlier multi-host evidence predates the final feed.
+
+### SQS store system instance (not wired)
+
+Spinifex can build, install, launch, restart and give storage to a dedicated SQS store microVM, but nothing calls any of it.
+No SQS API, domain service, admin command or daemon start-up path creates a store, so none of this is in production use.
+
+The guest image is built by `make build-sqs-store-microvm-image`, separate from the load balancer image, and bundles the `nats-server` version pinned in `go.mod` plus the ext4 tools.
+Release tarballs carry it as `microvm-sqs-store/`, and `setup.sh` installs it to `/usr/share/spinifex/microvm/sqs-store/`.
+A release without a complete image removes any previously installed one, so a launch is then refused rather than booting a stale image.
+
+At boot the guest reads three firmware config values: the data disk's serial, an IPv4 address on one of its interfaces, and one NATS user NKey public key.
+It mounts exactly one matching disk at `/data` after `e2fsck -p`, refuses a missing, duplicate, unformatted or unrepairable disk, and never formats.
+JetStream keeps its data on `/data` with `sync_interval: always` (ADR-0011:S3), and the server listens only on that address and accepts only that key.
+On shutdown the store is stopped before `/data` is unmounted; a store that survives SIGKILL leaves every mount in place.
+
+A launch uses the versioned subject `system.LaunchSQSStore.v1`, which a daemon without this code does not serve, so it fails with no responder instead of booting another guest.
+The legacy `system.LaunchInstance` path refuses the store kind on both send and receive.
+Every input check runs before capacity is reserved, and the data volume must be an available system-account volume.
+It is mounted through the normal EBS path and detached, never deleted, on terminate, which is the protection standard EBS gives today and nothing stronger.
+CPU mitigations stay enabled for the store, unlike the load balancer.
+A successful launch means the VM is running, not that NATS is ready.
+
+The instance record keeps the store's non-secret guest settings so a restart on the node that launched it rewrites the firmware config files.
+A relaunch on any other node is refused, and cross-node recovery never claims a store.
+Both checks exist only in daemons with this code: an older daemon can still claim and relaunch a store elsewhere, and an explicit start served by another node claims the instance before its relaunch is refused.
+
+`domains/sqs/storevolume` prepares a fresh data volume, because the guest never formats.
+It checks for `mkfs.ext4` and `qemu-img` before allocating, creates the volume in `creating`, writes an empty ext4 image to it over NBD on the local node, unpublishes it, and only then makes it `available`.
+While a volume is `creating`, attach, delete, modify, force-detach and snapshot are refused.
+A volume is deleted only after a confirmed unpublish; any publish error, an unconfirmed unpublish or a failed release keeps it and reports its ID, with its state reported as unknown after a failed release.
+
+Evidence is local only: unit tests, real QEMU boots against disposable scratch disks, and one run of the preparer against the `qemunbdd` provider whose volume kept acknowledged writes across a reboot.
+Viperblock, volume encryption, the EC2 volume record path, multiple nodes and host reachability of the guest address are not tested.
+Credential issuance and rotation, store placement and NATS readiness checks do not exist, and there is no failover: a lost node leaves its store unavailable.
+
+### SQS worker guest (operator-launched)
+
+The shared SQS worker guest holds the SQS store's client credential, so that no platform component does (ADR-0011:S2).
+Only an operator activates it, with the super-admin `EnsureSQSWorker` action described below; nothing starts it at daemon startup, and no worker service runs after boot.
+
+`make build-sqs-worker-microvm-image` builds it with the store image's script and network `/init` under `SQS_MICROVM_PROFILE=worker`; it carries the static `sqs-worker` binary and the ext4 tools, no `nats-server`, and an empty `/data` mountpoint.
+Releases build it beside the store image and ship it as `microvm-sqs-worker`, which `setup.sh` installs to `/usr/share/spinifex/microvm/sqs-worker` with the store image's rules: a release without it removes an earlier one, and an incomplete one is refused.
+At boot `sqs-worker boot` reads three firmware config values under `opt/spinifex/sqs-worker/`: the worker disk's serial, its Region and the gateway URL; none is a first-store decision, because a static value survives a guest reset.
+A missing, empty or malformed value refuses before any disk is touched.
+
+It selects exactly one disk with that serial, runs `e2fsck -p`, mounts it at `/data` and then confirms from `/proc/self/mountinfo` that this disk, as a whole ext4 filesystem, is what is mounted there; an empty mountpoint, a foreign mount or a failed check or mount refuses, and the guest never formats.
+Only then does it ask the gateway for the current decision with `ClaimFirstStore`, signed by `internal/sqsworkergw` with credentials from the instance metadata service only (`gwsign.NewInstanceRole`, no environment, profile or static-key fallback), over TLS 1.3 that trusts only the cluster CA from `opt/spinifex/ca-cert` and never follows a redirect.
+Under that decision `workerkey.Acquire` creates the key on `/data/sqs-worker-key`, owner-only and synced, records only its public key and then uses it, or loads the recorded key; a missing or different key once one is recorded refuses and is never replaced.
+The VM can start before its launch binds the instance, so its first calls are legitimately `AccessDenied`; that denial creates nothing and is asked again, as are a lost reply and a `ServiceUnavailable` or `Throttling` refusal, all within one bound (60 attempts 5 seconds apart), after which a worker that was never bound, or was replaced, refuses the boot.
+A signature or other refusal, or an answer that is neither "create" nor one valid public key, refuses the boot at once.
+
+`domains/sqs/firststore` is the SQS domain's first-store decision: one record per Region in the `spinifex-sqs-first-store` KV bucket moves from `claimed` to `key-recorded` to `launch-intent` under compare-and-swap, bound to the worker's instance ID, claim ID and generation.
+A claim authorizes key creation only while no public key is recorded; once one is, every claim returns only that key, and a store launch intent needs it recorded first.
+An older generation, a different identity at the same generation or an unreadable record authorizes nothing, and the record holds no private key material.
+Reads are stream message gets answered by the stream leader, never direct gets a lagging replica can answer, and every authorizing answer is also confirmed by a compare-and-swap at the revision it read, so a stale or unavailable read authorizes nothing.
+The authenticated caller that will carry this decision must derive the worker's instance ID and generation from authoritative instance state; a generation supplied in a request cannot itself authorize a takeover.
+
+`domains/sqs/workerident` is that authoritative instance state: one record per Region in the `spinifex-sqs-worker-identity` KV bucket holds a monotonic generation, claim ID, pinned node and encrypted data volume, and moves from `launching` to `bound` to `stopped`, or from `launching` to `failed`, under compare-and-swap.
+A launch intent is refused while the previous attempt is `launching` or `bound`, including one whose outcome is unknown, so only one caller ever holds an intent to start a VM on the pinned volume.
+A replacement takes the next generation only after `FailLaunch` or `RecordStopped` records that nothing holds the volume, and each request names the generation it supersedes, so a replayed request cannot take a generation it never observed; there is no automatic takeover.
+The launch path then raises the first-store fence to the successor's generation with `firststore.Fence` before `BeginLaunch`, which refuses a replacement until a leader read shows that fence; the fence only rises, and `firststore` refuses every call from a generation below it.
+Only the current claim ID is recorded, so claim IDs are minted fresh per attempt and a VM launch must be idempotent on generation and claim together.
+Binding writes the instance's index entry first and then binds the record to that instance.
+The index is only a lookup: `Authorize` accepts an instance only while the Region's record is `bound` to exactly that instance, generation and claim, confirmed by a compare-and-swap at the leader-read revision, so an old, orphaned, rewritten or unbound entry authorizes nothing.
+A same-instance reboot keeps its instance ID and identity; a replacement takes a new generation, and the first cut refuses another node or data volume and an unencrypted volume, although the encryption flag is the caller's claim and the launch wiring must verify it from the volume's metadata.
+Both buckets are opened through `OpenDurableKV`, which refuses an existing bucket whose effective stream configuration could expire, evict, lose, lag or refuse records: non-limits retention, an age, per-message or delete-marker TTL, memory storage, asynchronous flush, a size, count or message-size cap, a discard policy other than discard-new, unacknowledged writes, a sealed stream, a mirror or sources, or fewer replicas than the cluster needs.
+`workerkey.Acquire` drives that decision: it syncs a new key to disk, records its public key and only then returns it, so a crash before recording reuses the key already on disk.
+
+`domains/sqs/awsapi/worker` carries the decision to the worker as `ClaimFirstStore` and `RecordWorkerPublicKey`, served under the internal `spinifex-sqs` signing service, which no AWS SDK signs for; the public AmazonSQS action inventory is untouched.
+The gateway admits only an assumed-role session of `sqsWorkerRole` in the system account, then the role's policy, and only then resolves the session name, the instance ID STS set, through `workerident.Authorize`; the request carries no identity, and any parameter beyond the action's own is refused.
+The first-store identity is the current record's instance, claim and generation, and a worker bound to another Region is refused.
+A request that passed `Authorize` just before a confirmed stop and replacement still carries the old generation; the fence was raised before the successor began, so `firststore` refuses it whether or not the successor has claimed yet.
+The adapter never logs a caller-chosen parameter name, because one could carry a secret.
+`RecordWorkerPublicKey` accepts only a valid public user key, so a seed is refused without being recorded, logged or echoed.
+The gateway binds both buckets per request and never creates them, so it provisions nothing and answers `ServiceUnavailable` until a worker launch has created them; the worker launch below ensures the role.
+The private key stays on the worker disk; the guest logs and publishes only the public key to `/run/sqs-worker/public-key`.
+
+`domains/sqs/workerlaunch` realizes the worker on the node it runs on, its only node: a bound record is returned as is; otherwise it takes the intent, then on the first launch marks the intent's one volume create, prepares the volume and records its ID, reads the volume's own record and refuses it unless encrypted, ensures the role, finds or creates the network for the claim and reads the recorded store key, and marks the intent dispatched before launching.
+A dispatched launch is never sent again: a lost reply or error leaves the intent `launching` and dispatched, and every later call returns `ErrOutcomeUnknown` until the outcome is reconciled, since the launch path is not idempotent.
+A failure before dispatch leaves an undispatched intent that the next call resumes with the same claim; a confirmed stop or failure is replaced only after `firststore.Fence` excludes the old generation, reusing the pinned volume.
+The launch passes no store decision or key: the guest asks the gateway at boot.
+The intent exists before any volume, so concurrent first launches allocate nothing for the losers. A Prepare proven to have sent no create withdraws the mark and returns `ServiceUnavailable`, or `UnsupportedOperation` when the cluster storage key is not configured, and a retry may prepare the volume; a create that may have applied is never sent again, and every later call returns `IncorrectState` until it is reconciled. A volume the call knows of but cannot use is reported with its ID in a `VolumeError` and left in place, never deleted. An intent that has sent no create may be re-typed; one that has keeps its type until replaced. The worker role is created with EC2-only trust, and an existing role is checked for that trust before any policy is granted; the instance profile must carry exactly the worker role, and one holding any other role is refused unchanged.
+The daemon's `LaunchSQSWorkerInstance` boots the `sqs-worker` image through the same direct-boot path as the store, with the data volume attached at boot and kept on terminate, a required instance profile, and a required mgmt NIC; it re-reads the volume's own record and refuses one that is not available and encrypted, before anything is allocated.
+Its firmware config adds the gateway URL on this node's mgmt bridge and the cluster CA from the daemon's own trust store; the caller cannot supply either, and no value written is secret.
+A relaunch rebuilds that config from the VM record on the node that launched it and is refused anywhere else, and membership takeover already skips SQS-managed instances.
+`sqsWorkerRole` is ensured in the system account granting exactly `spinifex-sqs:ClaimFirstStore` and `spinifex-sqs:RecordWorkerPublicKey`, from the `domains/sqs/workerrole` contract the gateway gate also reads; an existing role whose trust admits more than the EC2 service is refused.
+
+Evidence is local only: unit tests with a fake firmware config, sysfs and mount table, tests running the boot through the real client against the real `spinifex-sqs` registration and records, and QEMU boots of the built image against disposable scratch disks and a host-side test gateway and metadata service, showing the first boot creates the key, reboots reuse it, a key lost after recording is never replaced, and the seed reaches no log or request.
+The worker's network is a private SQS system VPC per Region, built by `systemvpc.EnsurePrivate` from the `[sqs]` supernet (default `10.240.0.0/14`): one VPC and one /24 subnet, with no internet gateway, NAT gateway, EIP, route table or daemon-to-store node port. Each launch claim gets one ENI in that subnet, in the VPC default security group, found again by its claim on a retry; two ENIs for one claim are refused. The launch intent is marked before the one create is sent, so a create whose ENI is not yet visible is never sent again: the launch fails closed with `IncorrectState` until the ENI appears or the outcome is reconciled. Worker-to-store TCP and the guest's route to the metadata service are not yet proven by any traffic test.
+The super-admin `EnsureSQSWorker` action on the internal `spinifex` operator service takes a `NodeId`, which must be a node of this cluster and a single subject token, and an `InstanceType`, and sends a request to `sqs.worker.Ensure.v1.<node>` with no queue group, so only that node answers and the worker stays pinned.
+That node's daemon re-checks the node the request names and the instance type, and refuses as `ServiceUnavailable` before allocating a volume or network or recording an intent unless the worker image's kernel and initramfs are non-empty files, its mgmt bridge exists and the cluster CA is readable; a call after the image is installed then launches.
+It then runs its one composed launcher for its own Region, one launch at a time, in the daemon's own context with a 30-minute bound, so a caller's timeout does not cancel it.
+A reply lost to that timeout is reported as `ServiceUnavailable`; calling again reports the bound worker, or `IncorrectState` for a dispatched launch whose outcome is unknown, which is never resent; a Region pinned to another node is refused as `InvalidParameterValue`.
+The reply names the bound VM's node, Region, instance ID and generation; bound is not ready, and it says nothing about the guest's boot, its key or the store.
+Stop/detach evidence for `RecordStopped` is not wired, the guest's metadata route and gateway TLS on a real ENI and worker-to-store TCP are unproven, and Viperblock's at-rest key is held by the platform.
+
+The worker identity record holds the instance type of its generation; a request for another type is refused as `IncorrectState`, before dispatch and after binding, the reply names the recorded type, and only a replacement after a confirmed stop may change it.
+Operators call both actions with `spx admin sqs ensure-worker --node --instance-type` and `spx admin sqs ensure-store --node --instance-type --volume-size-gib`, signed for the `spinifex` service with the standard AWS credential chain like the other `spx admin` gateway commands.
+Each command refuses missing inputs before sending anything and reports only the bound VM; a bound VM is not a ready worker or store.
+The super-admin `EnsureSQSStore` action takes a `NodeId`, an `InstanceType` and a `VolumeSizeGiB` and sends a request to `sqs.store.Ensure.v1.<node>` with no queue group; only the node the Region's worker is bound to can launch the store.
+That node refuses an instance type whose memory does not exceed 512 MiB, the store's `GOMEMLIMIT` under ADR-0011:S11; this is a floor, not a validated size, and no type is defaulted.
+It refuses before recording or allocating anything unless the store image and the cluster storage key are present, the worker is bound on this node, and that worker's identity has recorded its public key in the first-store decision, which it then advances to a launch intent.
+The store's own launch record in `spinifex-sqs-store-launch` holds only that public key, the claim, the node and the operator's instance type and size, and marks each step before its uncertain create: the volume, the ENI in the worker's private subnet, and the launch.
+A later request for a different instance type or size is refused as `IncorrectState`, before dispatch and after binding, and the reply names the type the store was launched with.
+A volume preparation that failed before sending any create, such as a missing `mkfs.ext4` or a scratch or format failure, withdraws its mark and returns `ServiceUnavailable`, so a retry may try again once the host is fixed.
+Any other marked step whose result is not visible is reported as `IncorrectState` and never resent; a volume that is not encrypted by its own record is refused as `UnsupportedOperation` with its ID kept, and the daemon re-checks the volume's encryption and the guest's size at launch.
+The daemon never receives the worker's seed; the store's NATS server accepts only that public key, and the reply's bound VM and address are not readiness, which only the worker can check.
+
+After boot, the worker's long-running `sqs-worker ready` step mounts and verifies its disk again and calls the internal `spinifex-sqs` `DescribeStore` action, which admits only the current worker and returns the bound store's address and recorded public key, or pending.
+It loads its key only if it matches that public key, then connects to the store's NATS on port 4222 with it and asks JetStream for account info, each pass bounded to 30 seconds.
+The step runs until shutdown with no overall deadline, so a store the operator launches at any time is picked up: it asks again every 5 seconds while the store is pending, not answering or the reply was lost, and re-checks a ready store every 30 seconds.
+A pass writes only the address and public key to `/run/sqs-worker/store-ready`; any failed pass removes that file, and so does the step's own exit.
+A start that finds shutdown in progress or no worker key logs once and waits in place rather than exiting, so init does not respawn it in a loop; shutdown stops it like a running watcher.
+Init respawns the step, at most once every 5 seconds, and every start removes a marker left by a watcher that was killed or crashed before the new watcher runs.
+A ready store that fails is noticed up to one re-check interval plus one bounded call later, so the marker can be up to about a minute stale.
+Shutdown blocks further respawns, removes the marker, and stops the watcher by its recorded PID and start time, escalating to SIGKILL after a budget; it unmounts the worker disk only once the watcher is confirmed gone, and a survivor leaves every mount in place.
+A mismatched key, a store that refuses the key, a key missing from the disk or an unclassified gateway refusal is asked again only every 5 minutes, so it never spins and still recovers if the operator fixes it; a missing key is never replaced.
+The NKey authenticates the worker to the store, but nothing authenticates the store to the worker: there is no TLS, so a pass means only that something at that address accepted the key.
+This is store readiness for the worker alone; there is no SQS API, and no production sizing or support is claimed.
+
 ## Configuration

 Cluster configuration (`spinifex/bootstrap/config/config.go`):
@@ -351,8 +351,24 @@ while the branch is in flight; they do not describe the umbrella repository's
 | `spinifex#1269` | `spinifex/handlers/elbv2` | `spinifex/domains/elbv2` (package `elbv2`) | A mechanical whole-package relocation under the 2026-10-26 ADR-0001 amendment: the ELBv2 control plane keeps its files, tests and current dependencies, and importers alias it `elbv2domain`. No identifier, API, NATS subject, KV bucket, metric, error string or behaviour changes. Its allowlist entries are now sourced from `domains/elbv2`. Q-30's data-plane boundary stays deferred, recorded as `agents/elbv2` (Q-30 data-plane boundary) debt. |
 | `spinifex#1271` | `spinifex/handlers/ecs` | `spinifex/domains/ecs` (package `ecs`) | A mechanical whole-package relocation: ADR-0001's `handlers/ecs` row targets `domains/ecs`, the files are already organised by resource, and resource packages are introduced only where callers need a compiler boundary. The ECS control plane keeps its files, tests and current dependencies alongside the already-extracted `domains/ecs/taskdefinition`, and importers alias it `ecsdomain`. No identifier, API, NATS subject, KV bucket, metric, error string or behaviour changes. Its allowlist entries are now sourced from `domains/ecs`. `handlers` has no package left and is a retired legacy root; R5 holds `domains/ecs` to the interim rule's ban on importing the daemon. |
 | `spinifex#1273` | `spinifex/gateway/{ecs,elbv2}` | `spinifex/domains/{ecs,elbv2}/awsapi` (`awsapi`, imported as `ecsawsapi`, `elbv2awsapi`) | A mechanical relocation of the ECS and ELBv2 AWS action adapters to their ADR-0001 domain homes, as `spinifex#1265` did for IAM, STS, EKS and RDS; APIs, NATS subjects, KV buckets and behaviour are unchanged. The root `gateway` dispatch still imports each package, recorded as `gateway` (service registration) debt, and `domains/admission/quota` keeps its live ELBv2 counter import, now of `domains/elbv2/awsapi`. |
+| `spinifex#1336` | `spinifex/gateway/tagging.go` (`Tagging_Request`, `taggingActions`, `taggingActionFromTarget`) and the `GatewayConfig.TaggingResources` field | `spinifex/domains/tagging/awsapi` (`NewRegistration`, `Deps`, `Actions`, `OperationInventory`), composed by `spinifex/runtime/roles/awsgw` | Resource Groups Tagging registers through the ADR-0002 seam like ACM and Route 53. The check order is kept: action, scope, authorization, account, body, aggregator, so a denied caller still gets `AccessDenied` before any body error and an unwired aggregator answers `ServerInternal`. The `checkPolicyResources` stand-in of `*` for every action is unchanged. The legacy switch case, the `supportedServices` entry and the `jsonErrorService` case are removed, and the registration declares the JSON 1.1 envelope. No allowlist row existed for this path. The nil-NATS check moves to the gateway-wide gate, which answers `ServiceUnavailableException` before dispatch as it did for every service. `cmd/aws-model-coverage` output is byte-identical because there is no embedded Tagging model. The characterization tests that pin the pipeline behaviour were committed first and pass unchanged against the registration; only the `withTagging` test helper changed. |
+| `spinifex#1336` | `spinifex/gateway/elbv2.go` (`ELBv2_Request`, `elbv2Actions`, `elbv2Handler`, `accountLoadBalancerLimit`) | `spinifex/domains/elbv2/awsapi` (`NewRegistration`, `Deps`, `Quota`, `ActionNames`, `OperationInventory`), `spinifex/domains/admission/quota` (`Service.LoadBalancerLimit`), `spinifex/ingress/aws/query` (`ParseArgs`), composed by `spinifex/runtime/roles/awsgw` | ELBv2 registers through the ADR-0002 seam as the first AWS Query service. The check order is kept: query parse, action, account, input parse, scope, authorization, NATS, dispatch, so a denied caller still gets `AccessDenied` before the handler runs. The gateway members ELBv2 read are now supplied as `Deps`: `NATSConn` as `Deps.NATS`, `Quota` as the `Deps.Quota` interface declared in `awsapi` because `domains/admission/quota` imports `awsapi`, `Region` as `Invocation.Region`, and the policy gate as `Invocation.Authorize`. The `DescribeAccountLimits` cap resolution moves to `Service.LoadBalancerLimit` unchanged, and a nil `Deps.Quota` enforces nothing and reports the AWS default. AWSGW now wires the registry after it builds the quota service, because the registration captures it. The registration declares the XML envelope and the inventory name `elasticloadbalancingv2`, so `cmd/aws-model-coverage` keys it explicitly and its output is byte-identical. `gateway.ParseAWSQueryArgs` delegates to `query.ParseArgs`. The legacy switch case, the `supportedServices` entry, the inventory entry and the `gateway -> domains/elbv2/awsapi` allowlist row are removed. `cmd/aws-model-coverage` imports `domains/elbv2/awsapi` for the inventory, under a fourth named-exception allowlist row whose reason is now generic for all four. Finding, not fixed here: the XML for `DescribeAccountLimits` and `DescribeLoadBalancers` lists member fields in a different order from run to run because the marshal is map-ordered, so the characterization tests pin decoded content rather than bytes. The characterization tests that pin the pipeline behaviour were committed first and pass unchanged against the registration; only the `withELBv2` and `serveELBv2` test helpers and the setup of two older gateway tests changed. |
+| `spinifex#1337` | `spinifex/gateway/spinifex.go` (`Spinifex_Request`, `spinifexAdminActions`) | `spinifex/operator/admin` (`NewRegistration`, `Deps`, `ModelAccess`, `ActionNames`), composed by `spinifex/runtime/roles/awsgw` | The `spinifex` Query service registers through the ADR-0002 seam with the XML error envelope it had. The check order is kept: query parse, action, authorization with the wildcard resource, account, the super-admin account gate, then the action, so a denied caller learns nothing about which actions exist. The gateway members the actions read are now `Deps`: `Version` and `Commit` strings, `NATSConn` as `Deps.NATS`, `RootCAs`, `DiscoverActiveNodes` as `Deps.ActiveNodes`, `BedrockAccessAdmin` as the `ModelAccess` interface, and the credential and access resolvers as `Deps.Credentials` and `Deps.Access`. The query actions need only the caller account, so the actor is not widened. The cluster-unavailable answer stays in the shared request path ahead of authorization. `/admin/{method}` moved onto the same registration in `mulga-3nbla.7`, recorded below. |
+| `spinifex#1344` | `spinifex/gateway/rds.go` (`RDS_Request`, `rdsCaller`) | `spinifex/domains/rds/awsapi` (`NewRegistration`, `Deps`, `Quota`, `CallerFromActor`, `OperationInventory`), composed by `spinifex/runtime/roles/awsgw` | RDS registers through the ADR-0002 seam with the XML envelope it had. The check order is kept: query parse, action, account, the principal-class gate for the internal agent actions, scope, authorization, NATS, quota and dispatch, so a denied caller still gets `AccessDenied` before the handler runs and no customer grant is evaluated against an agent action. The gateway members RDS read are now supplied as `Deps`: `NATSConn` as `Deps.NATS`, `ExpectedNodes` as `Deps.ExpectedNodes`, `Quota.EnforceRDSInstances` as the `Deps.Quota` interface, `Region` as `Invocation.Region` and the policy gate as `Invocation.Authorize`. A nil `Deps.Quota` skips the DB-instance cap. The caller comes from `dispatch.Actor`: the session name for an assumed-role session and the user name otherwise, and the role name from `Actor.RoleName()`, which parses the underlying role ARN and never the session name. The account comes from `Invocation.AccountID`, so a caller with no composable ARN still reaches the policy check and is refused there as before. The legacy switch case, the `supportedServices` entry, the inventory entry and the `gateway -> domains/rds/awsapi` allowlist row are removed, and `cmd/aws-model-coverage` reads the inventory under a fifth named-exception allowlist row. The characterization tests that pin the pipeline behaviour were committed first with spinifex#1336 and pass unchanged against the registration; only the `withRDS` and `serveRDS` test helpers and the setup of the direct-call RDS tests changed. |
 | `#1277` | `spinifex/accountteardown` | `spinifex/operator/accountlifecycle` (package `accountlifecycle`, imported as `accountteardown`) | A mechanical whole-package relocation under the 2026-10-27 ADR-0001 amendment: the account deletion coordinator keeps its APIs, NATS subjects, KV access and behaviour; importers keep the `accountteardown` name as an alias, so only their import paths change. Replacing its direct domain and adapter cleanup calls with owner capabilities stays deferred debt. |
 | `#1285` | `spinifex/gateway/requestaudit.go` and `slogRequestLogger` from `spinifex/gateway/gateway.go` | `spinifex/ingress/aws/audit` (package `audit`) | The first generic-ingress extraction from the root gateway: the request-audit middleware and access-log middleware move unchanged in behaviour. The gateway's authentication flow and dispatch still populate the record, now through the exported `audit.From` and `Record` setters at the same call sites; the package imports only `foundation/telemetry` and `ingress/http`. Its unit tests moved with it; the tests that run it behind SigV4 stay in `gateway`. |
+| `spinifex#1334` | `spinifex/gateway/spx` (`spx.go`, `storage.go`, `promote_image.go` and their tests) | `spinifex/operator/admin` (package `admin`) | The version, node, VM and storage-status gather helpers and the image-promote NATS request move unchanged with no shim; the `spinifex` service dispatch that calls them stays in `gateway/spinifex.go` until the `spinifex` registration slice. `operator/admin` already imports `domains/ec2/image` and `contracts/operator/v1`, so the move adds no edge and the layering ratchet needed no allowlist row. The `gateway/spx` frozen-legacy entry is removed. |
+| `spinifex#1334` | `spinifex/gateway` (`CreateAccountRequest`/`Response`, `DeleteAccountRequest`, `DescribeAccountDeletionRequest`, `DeletionState*`, `AccountSummary`, `ListAccountsResponse`, `AdminMethodNames`) | `contracts/operator/v1` (`admin.go`) | The `/admin` bodies the CLI shares with the gateway move unchanged in field names, JSON tags and omitempty, pinned by golden JSON written against the gateway types first and moved with them. `AdminMethodNames` is now owned by the contract and the gateway builds its `adminMethods` accept set from it, with a test that the router accepts exactly that list. `AccountQuotaRequest`/`AccountQuotaResponse` (embed `quota.Overrides` from `domains/admission/quota`) and `DeleteAccountResponse` (embeds `accountlifecycle.Result`) stay in `gateway`: `contracts` may import only `foundation` and `contracts`, so they are settled when the `/admin` handlers move to `operator/admin`. `accountlifecycle.Stage` serialises as an integer, not its name; the golden test pins that. |
+| `spinifex#1345` | `spinifex/gateway/admin*.go` (the `/admin/{method}` handlers, gates and body cap) and the `AccountQuotaRequest`/`AccountQuotaResponse`/`DeleteAccountResponse` wire types | `spinifex/operator/admin` (`rest.go`, `account_*.go`), served by the same `spinifex` registration | `/admin/{method}` stays a chi route in `gateway`, ahead of the service selector, but `Admin_Request` now only mints the request ID and header and hands the request to the registered entry through `dispatchRegistered`, the path every Query request takes. `dispatch.Invocation` gains `Service`, the credential-scope service, so the scope gate is the first identity gate in the dispatcher and a key signed for another service is denied exactly as before. The order is unchanged: verb and method, then the five fail-closed gates (scope, super-admin account, IAM user, identity, `spinifex:<Method>` policy), then the cluster check, then the 8 KiB body cap. The dispatcher returns errors and the route renders them as the JSON envelope with `requestId`, then logs once from the real result, so a denial or a 5xx is never logged as a completed request; the registration's `ResolveAction` labels the method, which removes the `auth.go` special case. The caller is read from `dispatch.Actor` (`UserName` for the gate and `updatedBy`). `BucketStore` and `SignupMaxAccounts` leave `GatewayConfig` for `operator/admin.Deps`, composed in `awsgw` beside `IAM`, `Quota`, `ExpectedNodes` and `NodeIDs`. The three wire types that could not enter `contracts` move to `operator/admin` with their handlers, `operator/cli` imports them from there, and the last `operator/cli` → `gateway` allowlist entry is deleted. |
+| `spinifex#1346` | `spinifex/gateway/iam.go` (`IAM_Request`, the IAM action table, `getUserForCaller`, the PassRole check) | `spinifex/domains/iam/awsapi` (`NewRegistration`, `Deps`, `ActionNames`), composed by `spinifex/runtime/roles/awsgw` | IAM registers through the ADR-0002 seam with the XML error envelope it had. The check order is kept: query parse, action, service and account checks, then one parse of the input, the `GetUser` default for a missing `UserName`, scope resolution, authorization, the `iam:PassRole` check for `AddRoleToInstanceProfile`, then the handler. The caller is read from `dispatch.Actor` (`PrincipalType`, `UserName`), with no ARN parsing. The dependencies are `Deps.IAM` and `Deps.CountInstanceProfileAssociations`, composed in `awsgw`, and the gateway keeps no IAM fallback. The `cmd/aws-model-coverage` named exception gains an `iam` row. |
+| `spinifex#1346` | `spinifex/gateway/sts.go` (`STS_Request`, the STS action table, the anonymous STS dispatch) | `spinifex/domains/sts/awsapi` (`NewRegistration`, `Deps`, `ActionNames`, `NewAnonymousAssumeRoleWithWebIdentity`), composed by `spinifex/runtime/roles/awsgw` | STS registers through the ADR-0002 seam with the XML error envelope it had. The check order is kept: query parse, action, service check, caller, the `AssumeRole` identity gate scoped to the stored role ARN, then the handler. The caller is read from `dispatch.Actor` (`CallerARN`, `UserName`, `SessionName`, `UserID`, `AccessKeyID`), with no ARN parsing. The gateway's authorization failure wraps a generic `dispatch.PolicyDenialError`, and STS matches it to answer with the caller, action and resource, so the gateway holds no STS-specific denial code. An unsigned `AssumeRoleWithWebIdentity` stays ahead of SigV4 in the gateway and reaches the handler through `GatewayConfig.AnonymousSTS`, a func composed in `awsgw`, never the STS service. The `cmd/aws-model-coverage` named exception gains an `sts` row. |
+| `spinifex#1348` | `spinifex/gateway/ecs.go` (`ECS_Request`) and `spinifex/gateway/eks.go` (`EKS_Request`, the EKS route table) | `spinifex/domains/ecs/awsapi` and `spinifex/domains/eks/awsapi` (`NewRegistration`, `Deps`, `OperationInventory`), composed by `spinifex/runtime/roles/awsgw` | ECS (AWS JSON 1.1) and EKS (REST JSON) register through the ADR-0002 seam with the JSON error envelopes they had, both declared as AWS JSON 1.1 (see the envelope divergence below). The check order and every response are kept, pinned by the pipeline characterisation tests committed first. EKS and ECS resolve roles through `foundation/aws/passrole` `Check` with `domains/iam` `RoleResolver`: a malformed or foreign role is refused first, a missing role is authorized as written before it is reported, and a resolver outage is returned unchanged. Role handling now follows the AWS capture of 2026-10-30. ECS RegisterTaskDefinition refuses a foreign role with `ClientException` "Role is not valid" before PassRole and accepts a nonexistent, invented-path or non-ARN role from a caller who can pass roles. ECS RunTask refuses a foreign `taskRoleArn` or `executionRoleArn` override the same way and answers a nonexistent or invented-path one with "ECS was unable to assume the role". EKS CreateCluster and CreateNodegroup answer a non-ARN or foreign role with `AccessDeniedException` "Cross-account pass role is not allowed.". EKS public routes now report every parameter fault as `InvalidParameterException`, the code the EKS model declares, while the internal control-plane routes keep `InvalidParameterValue` because the guest agents match on it. Two deliberate EKS divergences remain. First, CreateCluster still refuses a nonexistent or invented-path role with `InvalidParameterException` "role ... does not exist", where AWS accepts it and the cluster later ends FAILED. Our EKS only stores, validates and echoes the cluster role and never assumes it, so accepting would leave an ACTIVE cluster bound to a role that does not exist. Second, CreateNodegroup keeps its existing refusal of a nonexistent role because AWS was not captured for it. A third divergence is the error envelope. EKS is restJson1 but registers `ErrorEnvelopeJSON`, so errors carry `application/x-amz-json-1.1` and a `__type` body where AWS sends `application/json` and a `message` body. The `X-Amzn-Errortype` header is correct and SDKs read the code from it, so impact is low. Switching to `ErrorEnvelopeRESTJSON` also changes the throttle code and the default internal-error code (`ServiceException` where the EKS model declares `ServerException`), so it is left for its own change. Verified live on env14: the ECS and EKS role cases above, and a valid CreateCluster reaching ACTIVE with the control-plane agent publishing through the gateway. The EKS route table and its tests move with the registration, and the gateway keeps no ECS or EKS fallback. The `cmd/aws-model-coverage` named exception gains an `ecs` and an `eks` row, and the two `gateway` allowlist rows are removed. |
+| `spinifex#1350` | `spinifex/gateway` (`ecr.go`, `ecrauth_middleware.go`, the `/v2/token` exchange, `ecr_principal.go` with `ResolveECRPrincipal`, the per-operation `/v2` authorization, the registry host matcher, `ctxECRPrincipal`) | `spinifex/domains/ecr/registry` (`Bridge`, `NewBridge`, `BridgeConfig`, and its own `Principal`, `Rehydrator`, `Authorizer`, `Admission` types with `ErrInvalidCredential` and `ErrUnavailable`), `spinifex/ingress/http` (`HostMatcher`, `TargetAccount`, `RequestClientIP`; names no service), `spinifex/ingress/aws/credential` (`Rehydrator`, `Principal.CallerARN`), and under `spinifex/runtime/roles/awsgw`: `credverify` (`Rehydrate`, sharing the SigV4 key and session lookups), `policyauth`, `accountadmission`, `registryadapt` | ADR-0001 amendment 2026-10-29 and ADR-0005:S2: the OCI `/v2` bridge and its token auth move out of the gateway, and ECR resolves its principal through the credential capability instead of reading `IAMService` and `STSService` directly. ADR-0002:S5 keeps `/v2` outside the seam, so the gateway keeps only a `RegistryHandler` mount that answers 501 when unset. `domains/ecr/registry` imports no ingress package (`TestADR0001_Layering` unchanged): the principal, rehydrator and authorizer are registry-local types and `registryadapt` maps `credential.Failure` codes onto `ErrInvalidCredential` (401) and `ErrUnavailable` (503). No widening of `dispatch.Actor` or `dispatch.Invocation`. The SigV4 path and the bridge share one policy evaluator (`policyauth`) and one account-admission check. Deliberate change: the open-registry mode for a nil token verifier is deleted and `NewBridge` refuses a nil Verifier, Issuer, Rehydrator, Authorizer or Admission, so tests that want an unauthenticated registry mount `registry.Handler` directly. The pre-move pipeline characterisation (status codes, error bodies, the `WWW-Authenticate` challenge, referrers, token exchange, revoked and deleted principals) passes unchanged apart from setup, and the retired `ResolveECRPrincipal` parity rows live on as `credverify` `Rehydrate` rows and `Bridge` claim-check rows. Live on env14 (DHCP external IP): crane, skopeo and docker push and pull (`tests/e2e/ecr`), a cosign sign and verify in the legacy tag mode and in OCI 1.1 referrers mode with the signature returned by `GET /v2/{repo}/referrers/{digest}`, `/v2/token` refresh from fresh logins, and a deactivated access key answering 401 on `/v2/token`, manifest GET, crane pull and crane push, then 200 again once reactivated. The ECR signing-key deadlock bead (`mulga-6wqce`) is untouched. |
+| `spinifex#1351` | `spinifex/gateway/{bedrock,bedrockruntime,bedrockagent,bedrockagentruntime}.go` (`Bedrock_Request`, `BedrockRuntime_Request`, `BedrockAgent_Request`, `BedrockAgentRuntime_Request`, the four route tables), `spinifex/domains/ochre/authz.go`, and the 13 `Bedrock*` fields of `GatewayConfig` | `spinifex/domains/ochre/awsapi` (`NewRegistration`, `Deps`, `Quota`, `OperationInventory`, `ServiceName`), composed by `spinifex/runtime/roles/awsgw` | The four Bedrock services register through the ADR-0002 seam as one registration, per ADR-0002:S1 (one per signing service) and the ADR-0001 amendment of 2026-10-06: the service is `bedrock` and `bedrock-runtime`, `bedrock-agent` and `bedrock-agent-runtime` are aliases. A request signed with a sub-service scope is dispatched to that sub-service whatever its path, so a path owned by another sub-service answers `InvalidActionException` as before; only the shared `bedrock` scope is split by path. The pipeline characterisation tests committed first (status, error type and envelope per service, metering order, nil-capability answers) pass unchanged apart from setup. The 13 gateway fields become `awsapi.Deps`: `Credentials`, `Endpoints`, `Access`, `Recorder`, `Logging`, `Provisioned`, `Guardrails`, `Embedder`, `KB`, `DataSources`, `Vector` and `Quota`, with the documented nil fallbacks (no grants denies every model, a nil store errors rather than panics); `BedrockAccessAdmin` had no reader outside one test and is gone, and `Deps.NATS` is dropped because the nil-NATS answer is given by the gateway-wide gate before dispatch. `Logging`, `Provisioned`, `Guardrails`, `KB` and `DataSources` are the concrete exported store types, a temporary edge until Ochre exposes capability interfaces for them. The wire-typed action functions stay in the `domains/ochre` root and the SDK-typed adapters move; `authz.go` moves with them, which removes the `domains/ochre` to `ingress/aws/bodyscope` allowlist row, the only allowlist change. The legacy switch cases, the `supportedServices` and `jsonErrorService` entries and the path-based sub-service rewrite in `GetService` are removed, so `ec2` is the only service left on the legacy table; the rewrite's rows are pinned in `awsapi` as `resolveSubservice` and dispatch tests. The registration declares the AWS JSON 1.1 envelope it had and `cmd/aws-model-coverage` is not wired, so no coverage allowlist row is added. Recorded divergences, not changed by this move: Bedrock is restJson1 in AWS but answers with the JSON 1.1 envelope (`application/x-amz-json-1.1`, `__type` body), and CreateGuardrail answers 200 where AWS answers 202; both are tracked by `mulga-xug8b`. No change to `dispatch.Actor` or `dispatch.Invocation`. |
+| `spinifex#1354` | `spinifex/gateway/ec2.go` (the EC2 action table, `ec2Handler`, `ec2Actions`, the dispatch pipeline), `ec2_idempotency.go`, `ec2_validation.go`, and the EC2 token-bucket fields of `GatewayConfig` | `spinifex/domains/ec2/awsapi` (`Handler`, `NewHandler`, `Deps`, `Quota`, `Dispatch`, `ActionNames`, `InputHasField`), composed by `spinifex/gateway/ec2.go` | Slice 9a of the EC2 move: the capabilities first, with no behaviour change. The EC2 action table, idempotency and validation move to `domains/ec2/awsapi` behind `Deps{NATS, IAM, Quota, ActiveNodes, ExpectedNodes, NodeIDs, AZ, Endpoint, InstanceStatus}`; the gateway keeps the composition and `EC2_Request`, still on the legacy switch until slice 9b registers EC2 through the ADR-0002 seam. `Quota` is declared in `awsapi` because `domains/admission/quota` imports the EC2 action packages and cannot be imported back, and the gateway wraps `quota.Service` to supply the NATS instance-type resolver its retype gate takes. The instance and spot-instance adapters take the two-method `InstanceProfiles` interface (`ResolveInstanceProfile`, `CanonicalResourceARN`) instead of `iamdomain.IAMService`, which pays off the `spotinstance` to `domains/admission/quota` and `domains/iam` allowlist rows. Allowlist changes are 20 deletions and no additions: the 18 `gateway` to `domains/ec2/awsapi/*` rows and those 2. The pipeline characterisation tests committed first (error envelopes, check order, gateway-local actions, idempotent replay, PassRole, ImportKeyPair padding, launch-template expansion before authorization, RunInstances instance-profile refusal order) pass unchanged. Package-local tests drive every routed action through `Dispatch` against a stub daemon. No change to `dispatch.Actor` or `dispatch.Invocation`. Live on env14: DescribeRegions, DescribeAvailabilityZones and DescribeAccountAttributes, an idempotent CreateVolume replay and its parameter mismatch, the DryRun and MaxResults errors, a no-policy user refused with UnauthorizedOperation (including on the gateway-local DescribeRegions), an assumed-role session allowed and refused per its policy, and one RunInstances to running then terminated. |
+| `spinifex#1355` | `spinifex/gateway/ec2.go` (`EC2_Request` in production, the legacy service switch, its `supportedServices` and JSON tables) and the `Quota` field of `GatewayConfig` | `spinifex/domains/ec2/awsapi` (`NewRegistration`, `OperationInventory`, `ServiceName`), composed by `spinifex/runtime/roles/awsgw` (`ec2.go`: `newEC2Registration`, `advertisedEndpoint`; `ec2quota`) | Slice 9b of the EC2 move: EC2 registers through the ADR-0002 seam as service `ec2` and the legacy gateway switch is empty. The EC2 query-protocol error envelope is kept through `ErrorEnvelopeXML`, and the policy denial match is now `*dispatch.PolicyDenialError`. The characterisation tests committed first also pin the Authorize service name and action, the trace span and audit action, the throttle key (account plus action) and the cluster-unavailable gate, which runs before policy and action lookup for gateway-local and NATS-backed actions alike. They and the 9a tests pass unchanged apart from setup: STS and EC2 test setups now register the services they sign as, and the two tests that claimed a legacy service name are removed because no legacy name remains. `EC2_Request` survives only as a test shim. Allowlist changes are two deletions (the `gateway` to `domains/ec2/awsapi` and `domains/ec2/awsapi/instance` rows) and one addition, the per-service model-coverage row `cmd/aws-model-coverage` to `domains/ec2/awsapi`. `startInstanceStatus` stays in `awsgw.go` because ADR-0007:S6 records the raw instance-state uses by file. No change to `dispatch.Actor` or `dispatch.Invocation`. Live on env14: DescribeRegions, DescribeAvailabilityZones and DescribeAccountAttributes, an idempotent CreateVolume replay and its parameter mismatch, the DryRun and InvalidVolume.NotFound errors, a no-policy user refused with UnauthorizedOperation (including on the gateway-local DescribeRegions), an assumed-role session allowed and refused per its policy, and one RunInstances to running then terminated. The gateway request logs carry `service=ec2` and the per-action label. Not exercised live: a throttled request (the default limits are not reachable from one client; the throttle key is pinned by the characterisation tests) and an agent-signed EC2 call. |
+| `spinifex#1363` | `spinifex/gateway` context keys (`ctxService`, `ctxRegion`, `ctxAction`, the principal and access key id keys, and the unread `ctxQueryArgs`), the duplicate `callerPrincipal` and `principalContext` types, `gateway/caller_arn.go` (`buildCallerARN`), and the direct `policyauth`, `accountadmission` and NATS-connection reads in the auth pipeline | `spinifex/ingress/aws/reqctx` (the request record), `credverify.CallerARN`, and the gateway-consumed capabilities `PolicyEvaluator`, `AccountAdmission` and `Readiness`, composed by `spinifex/runtime/roles/awsgw` (`clusterready`) | Slice 11a of the gateway removal: the request-context record and the ADR-0002:S8 capabilities, with no behaviour change. The record is a copy-on-write value under one context key; the principal is sealed once and a second `WithPrincipal` keeps the first caller. The layering rule `S7_RequestRecordPrivateToIngress` lets only `gateway` and `ingress/aws/*` import `reqctx`, with a fault-injection test, and slice 11c moves the package to `ingress/aws/internal/reqctx` and drops the rule. The caller ARN is built once in `credverify` (root cases kept) and filled on `Principal.CallerARN`. The policy evaluator, account admission (`credential.ErrAccountNotActive`, with `accountadmission.ErrNotActive` an alias) and cluster readiness (`Available() bool`) are interfaces the gateway consumes and fails closed on when nil. The step 1 characterisation tests (throttle against both readiness gates and the policy gate for ECS and IAM, `/admin` skipping `GetService` and the second gate, the nil-connection difference between the two gates, five error writers each minting a request id that differs from the audit record, and the credential-scope region reaching only the span) pass unchanged apart from setup. No allowlist rows change, and there is no change to `dispatch.Actor` or `dispatch.Invocation`. Debt for slice 11c: the test-only `defaultCapabilities` seam (a package variable set by a test `init` so about 60 gateway test files can build a config from `IAMService` alone), the nil-connection case of the early readiness gate, the remaining `NATSConn` readers (`DiscoverActiveNodes` and the field that `awsgw` and the integration harness set), and the `IAMService`, `STSService`, `Config`, `RootCAs` and `InternalSuffix` fields of `GatewayConfig`, which no production code reads. Live on env14: DescribeRegions, DescribeAvailabilityZones and DescribeAccountAttributes, an idempotent CreateVolume replay and its parameter mismatch, the DryRun and InvalidVolume.NotFound errors, a bad access key refused with InvalidClientTokenId on a Query service and InvalidClientTokenIdException on a JSON service, a no-policy user refused with UnauthorizedOperation, and an assumed-role session allowed and refused per its policy. Not exercised live: a suspended-account refusal, a throttled request and the NATS-down 503, which the step 1 tests pin. |
+| `spinifex#1334` | `spinifex/gateway/oidc_discovery.go` (`eksAccountKV`, the discovery-document and JWKS read and construction) | `spinifex/domains/eks/cluster` (`OIDCPublisher`, `OIDCDiscoveryDocument`), composed by `spinifex/runtime/roles/awsgw` | The EKS domain now reads its own account bucket and builds the per-cluster discovery document and JWKS responses (ADR-0005:S4), and the gateway only maps the unauthenticated `/oidc/eks/...` routes onto an `OIDCPublisher` interface it is handed. Responses are byte-identical, pinned by `oidc_discovery_current_behaviour_test.go`, whose assertions are unchanged. The gateway no longer imports `domains/eks/cluster` in production code. `STS` keeps its own `ClusterJWKSResolver` (`OIDCJWKSReader`), untouched. Recorded debt, not repaired here: a stored JWKS is served verbatim with 200 even when it is not valid JSON. |
 The EC2 contract row is intentionally different from the directory moves: it creates
 a compatibility boundary. `ec2.cmd.*` retains its deployed one-token NATS
 subscription and permission shape, while callers use
@@ -387,9 +403,9 @@ the canonical IAM/STS principal before dispatch, because ECR must not depend on
 gateway identity context or ARN construction.

 All implemented ECR control-plane actions now dispatch through explicit ECR
-capabilities. The remaining ECR-named gateway code is generic `/v2/*` route
-assembly, SigV4/registry authentication and policy enforcement; it remains in
-gateway rather than inverting those dependencies into the ECR domain.
+capabilities. The OCI `/v2` bridge, its token exchange and its principal
+resolution now live in `domains/ecr/registry`; the gateway keeps only the
+`RegistryHandler` mount, and `awsgw` composes the bridge.

 ### ACM relocation residuals

@@ -438,7 +454,7 @@ These residuals were recorded by individual slices and must be resolved once the
 - `runtime/roles/vpcd` (placement): the role still mixes process/role supervision with network-controller responsibilities. `lease_reaper.go` (orphaned external-pool DHCP lease sweep), `ovn_controller_watchdog.go` (OVN controller liveness and restart), `host_eip_loop.go` (host-EIP reconcile loop) and `lease_datapath_reconcile.go` (lease datapath reconcile) are periodic network-control loops, not role wiring; `vpcd.go` itself also carries bridge-mode detection and verification, chassis-list parsing, external-CIDR resolution, DHCP manager startup and lease/IP-change handling, and gateway/host-EIP allocator and neighbour flush/prime construction alongside `Service.Start`. The future placement question is whether these belong in `domains/network`. Removal condition: an accepted decision on network-controller placement, and extraction of those loops and helpers behind the role.
 - ECR control-plane registration: `awsapi.Deps.NATS` is a raw `*nats.Conn`, not a narrow capability. Every NATS-backed `Handler` passes it to `domains/ecr.NewNATSMetaStore`, which calls `natsmsg.NATSRequest`, so narrowing it means changing the exported `Handler` type, every handler and the meta store; it must not be broadened further meanwhile.
 - ACM control-plane registration: `domains/acm/awsapi.Deps.NATS` (`registration.go`) is likewise a raw `*nats.Conn` carried by every `Handler`, for the same reason; it must not be broadened further meanwhile.
- `cmd/aws-model-coverage` imports `ingress/aws/dispatch`, `domains/ecr/awsapi`, `domains/acm/awsapi` and `gateway` to read ECR's and ACM's declared inventories. It is a non-serving tool and the only named exception to the ADR-0002 composition invariant; it must not construct a registration.
+- `cmd/aws-model-coverage` imports `ingress/aws/dispatch`, `gateway` and the `awsapi` package of each registered service (`domains/ecr/awsapi`, `domains/acm/awsapi`, `domains/route53/awsapi`, `domains/elbv2/awsapi`, `domains/rds/awsapi`, `domains/ecs/awsapi`, `domains/eks/awsapi`) to read their declared inventories. It is a non-serving tool and the only named exception to the ADR-0002 composition invariant; it must not construct a registration.
 - Ingress pipeline (ADR-0002 phase 4): the request-audit record, the trace action enricher, the throttle key functions and the gateway request-context keys stay in `gateway` until the auth/IAM and request-context boundary is designed. SigV4 authentication writes the audit record and context keys from four places, so moving them now would export exactly the setters ADR-0002 says must not be exported solely to make a directory move possible.
 - NAT projection delivery is inconsistent and is preserved as found: EIP `AssociateAddress` sends `vpc.add-nat` unacknowledged (`AnnounceNATAdd`), skipping the 45s OVN barrier every other add path waits on, and `DeleteNetworkInterface` sends `vpc.delete-nat` unacknowledged (`AnnounceNATDelete`), skipping the ack and the ErrNoResponders retry. The reconciler's orphan sweep is the only backstop for both. Clears when those call sites use `AddNAT`/`AddNATBestEffort` and `RemoveNAT`, a reliability change of its own.
 - `vpc.add-system-egress` has a `vpcd` subscriber and a contract entry but no producer (its former `domains/ec2/eip` publisher is gone); `domains/eks/cluster` still publishes the delete. Either a producer is missing or the add route is dead; it needs an owner decision.
@@ -446,7 +462,7 @@ These residuals were recorded by individual slices and must be resolved once the
 - `foundation/messaging/nats/nats_trace.go` still names its OpenTelemetry tracer `github.com/mulgadc/spinifex/spinifex/utils`. It is an emitted instrumentation-scope name that dashboards may filter on, so renaming it is a telemetry change, not a path fix.
 - `runtime/compute/cache.New`: two construction-time panics carry a lint exemption pending a constructor that returns an error.
 - `domains/ec2/systeminstance`: managed-service parent metadata is only the `ManagedBy` tag value.
- `domains/ochre` (action adapters): since `spinifex#1266` the Bedrock action layer (handlers, action authorization, event-stream framing and the CRUD, catalogue and Converse/Invoke entry points) shares package `ochre` with the domain core, stores and inference engine, and imports `ingress/aws/bodyscope`. Moving it to `domains/ochre/awsapi` is not mechanical: the adapters reach package-private store internals (the guardrail and provisioned-throughput store handles and get/put/update methods), and adapter tests override the package-private model catalogue. It needs core capabilities first, then the move; the top-level `gateway/bedrock*.go` HTTP handlers additionally wait for the ingress registration seam.
+- `domains/ochre` (action adapters): cleared by the Bedrock registration in `spinifex#1351`. The SDK-typed adapters and `authz.go` moved to `domains/ochre/awsapi`, the `ingress/aws/bodyscope` allowlist row is deleted, and the adapters take the concrete exported Ochre stores through `awsapi.Deps` as a temporary edge until Ochre exposes capability interfaces for them.
 - `domains/network`: `reconcile/intent.go` builds desired network state by reading EC2's KV buckets and decoding EC2's private VPC, subnet, ENI, route-table, gateway, EIP, NAT and security-group records, and `reconcile/drift.go` watches those buckets. ADR-0001 names this pair explicitly: EC2 and network communicate through a capability/projection or versioned contract, never by one being the other's KV reader. Step 7 fix: an EC2-owned `domains/ec2/projection` exposing network intent with list and watch (it may stay KV-backed, so no data moves); a narrow intent interface declared by `reconcile`; wiring at the composition root; and the duplicated `PurposeEIP`/`PurposeIGWLRP` constants and mirrored `SGRule`/`SGEvent` folded into the projection or `contracts/ec2/v1`. The drift-watch coverage test and reconcile scenario tests are the evidence the projection must keep. `domains/network` also imports `runtime/compute/vm` and `runtime/host/command` (domain-to-runtime).
 - `domains/ec2/instance`: imports `runtime/compute/vm` and `runtime/compute/gpu` (domain-to-runtime) and `domains/dns` (cross-domain implementation import).
 - `domains/ec2/instance/state` (dependency): the vpcd IMDS local reader imports `runtime/compute/vm` to return `vm.VM`. Removal condition: that type moves into a domain-owned package that `runtime/compute/vm` consumes (runtime to domain is allowed).
@@ -456,14 +472,12 @@ These residuals were recorded by individual slices and must be resolved once the
 - `operator/installer/firstboot` (security): `Config.InstallCallback` is rendered into `spinifex-firstboot.sh`, which is world-readable (0755). The field is currently unused (`fireInstallCallback` reads `SPINIFEX_INSTALL_CALLBACK` from the environment instead), but any future wiring must not place secrets or tokenised callback URLs in the generated script; sensitive callback credentials should reach first boot through a protected channel such as a 0600 file or a credential store.
 - `operator/cli` (known security risk): `spx admin join` contacts the formation server with `InsecureSkipVerify`, sends the join token as a bearer credential and receives the cluster CA certificate and key, master key, Viperblock key and shared credentials, so an on-path attacker during formation can capture the token and the cluster root secrets or substitute a CA. This predates the move; lint surfaced it when the tree left `cmd/`. It is accepted for now and carries a line-level gosec annotation stating it is not a false positive. Removal condition: authenticate formation TLS with a pinned key or certificate delivered with the join material, or an out-of-band CA, with tests proving a mismatched server certificate fails before credentials are sent.
 - `domains/eks/cluster` (potential security risk): `ClusterMeta.ControlPlaneTemplate` (`domains/eks/cluster/service_impl.go`) clears the template's credentials but keeps `OIDCPrivateKeyPEM` and the k3s `JoinToken`, so both are persisted in the `eks-account-{acct}` cluster record, which the gateway also reads. Whether that record is encrypted at rest is unconfirmed; if it is not, a reader of the bucket can forge service-account tokens for the cluster's OIDC issuer and join nodes to it. This predates the refactor and is not fixed in it. Removal condition: the key and token are held by a dedicated secret store (a future KMS or Secrets Manager capability is the likely owner) and the cluster record keeps only a reference, with a test proving the persisted record contains neither secret.
- `operator/cli` (dependency): the command tree still imports domain and service implementations directly (`handlers/{iam,eks}`, `domains/admission/quota`, `gateway`, `domains/ochre`, `vpcd`, the Viperblock adapter and runtime roles). The ADR's `cmd/spinifex` current-to-target map row requires command packages to reach domains through public capabilities; splitting supported use cases into `operator/*` packages is later work.
+- `operator/cli` (dependency): the command tree still imports domain and service implementations directly (`handlers/{iam,eks}`, `domains/admission/quota`, `domains/ochre`, `vpcd`, the Viperblock adapter and runtime roles). The ADR's `cmd/spinifex` current-to-target map row requires command packages to reach domains through public capabilities; splitting supported use cases into `operator/*` packages is later work.
 - `providers/ebs/viperblock` → `runtime/compute/nbd` and `runtime/host/process` (dependency, moved unchanged from `services/viperblockd`): the adapter formats NBD endpoints and manages nbdkit processes through runtime packages. ADR-0001 does not list providers → runtime as forbidden and the layering test does not enforce it (same category as `providers/ebs/conformance` → `runtime/compute/nbd`), but a provider should not depend on node runtime. Clears when NBD endpoint naming has a provider-facing home (a candidate for `contracts/ebs/v1`) and process supervision is injected by the role. The adapter also imports `providers/objectstore`, a provider-to-provider edge that is allowed.
- `gateway/ec2.go` (registration): the EC2 dispatch map stays a `GatewayConfig` method in `gateway` and imports every `domains/ec2/awsapi/*` package; `gateway/ec2_idempotency.go` also imports `domains/ec2/awsapi/idem` (`KVBucket`, `TTL`, `TokenAndParams`) for the dispatch-time idempotency wrapper. ADR-0001 makes the domain own its action registration; moving the map waits on the generic ingress registration seam and must not be done as a package move.
- `gateway` (service registration): the root `gateway` dispatch for IAM, STS, EKS, RDS, ECS and ELBv2 imports `domains/{iam,sts,eks,rds,ecs,elbv2}/awsapi`, which only AWSGW may wire. It clears when each service registers through the ADR-0002 ingress registration seam; as with EC2, moving the dispatch must not be done as a package move.
+- `gateway` (service registration): the root `gateway` dispatch for IAM, STS, EKS, RDS and ECS imports `domains/{iam,sts,eks,rds,ecs}/awsapi`, which only AWSGW may wire. It clears when each service registers through the ADR-0002 ingress registration seam; as with EC2, moving the dispatch must not be done as a package move.
 - `operator/accountlifecycle` (dependency): its EC2 reapers call `domains/ec2/awsapi/*` action adapters directly, as they did under `gateway/ec2`. The amended ADR-0001 `accountteardown` row keeps replacing these with domain-owned cleanup capabilities as deferred debt; it was not a relocation prerequisite. It also holds a read-only consumer of the EC2 platform-inventory contract (`contracts/ec2/v1`, `ec2.v1.PlatformInventory`), served by `domains/ec2/accountinventory`; only the explicit dry run (`Engine.DryRun`) calls it, to log a shadow comparison, and the key-pair, placement-group and launch-template reapers still use the `awsapi` adapters. A domain-owned cleanup capability for the proven-safe subset of those three kinds (key pairs and empty placement groups no instance, spot request or live launch-template version names, and launch-template residue; never live launch templates) exists as `domains/ec2/accountcleanup` (a Go interface with injected owners and reference readers, no NATS subject); nothing calls it yet. It also holds the StageAccount settings-seal reaper, a consumer of `contracts/ec2/v1`'s `ec2.v1.SealAccountSettings` and per-node `ec2.v1.AccountSettingsSealCapability.<node>` subjects, served by `domains/ec2/account` through the daemon; the reaper reads no KV and imports no domain package, reaching both subjects only through NATS and the configured node set `NewClusterEngine` already receives.
 - `domains/admission/quota` (dependency, moved from `handlers/quota` unchanged): `live.go` imports `domains/elbv2/awsapi` and `domains/rds/instance` for its live ELBv2 and RDS counters; `records.go` imports `runtime/compute/vm` for the instance-record vCPU sweep; `live.go`, `vcpu.go` and `storage.go` call `domains/ec2/awsapi/{vpc,eip,volume}` describe adapters to count VPCs, subnets, addresses and volumes, as they did under `gateway/ec2`. The ADR's `handlers/quota` current-to-target map row replaces these with reader capabilities once their final locations exist.
- `domains/ec2/awsapi/instance` (dependency): imports `domains/iam` (`IAMService` in `DescribeInstances.go`, `RunInstances.go` and `IamInstanceProfileAssociation.go`; `InstanceProfile` in the latter two), `runtime/compute/cache` (`NodeState`, `NodeUnknown`, `NodeLive`) and `runtime/compute/vm` (`VM`), both in `status_synthesis.go`. The IAM edge needs an IAM capability; the runtime edges are domain-to-runtime and need the instance record and node liveness exposed through an EC2-owned projection.
- `domains/ec2/awsapi/spotinstance` (dependency): imports `domains/iam.IAMService` and `domains/admission/quota.Service` in `RequestSpotInstances.go`.
+- `domains/ec2/awsapi/instance` (dependency): imports `domains/iam` for the `InstanceProfile` type that `InstanceProfiles.ResolveInstanceProfile` returns (`RunInstances.go`, `IamInstanceProfileAssociation.go`), `runtime/compute/cache` (`NodeState`, `NodeUnknown`, `NodeLive`) and `runtime/compute/vm` (`VM`), both in `status_synthesis.go`. The IAM edge is now the two-method `InstanceProfiles` interface plus that type; the runtime edges are domain-to-runtime and need the instance record and node liveness exposed through an EC2-owned projection.
 - Instance adapter consumers outside EC2: `domains/rds/awsapi/enginecatalog.go`, `runtime/daemon/eks_cp_control.go`, `runtime/daemon/rds_deps.go`, `domains/admission/quota`'s `vcpu.go` and `gateway/iam.go` (`CountInstanceProfileAssociations`) call `domains/ec2/awsapi/instance` directly; each needs an EC2 capability rather than the AWS action adapter.
 - `runtime/daemon` (Q-28 ownership and deployment): the daemon moved whole, keeping its current ownership, deployment and `daemon → gateway → handlers` dependencies. Q-28's write-side consolidation and any daemon decomposition or split remain deferred by ADR-0001 and are not paid by the move.

@@ -673,11 +687,11 @@ series above; "needs decision" is open.
 | `handlers/elbv2` | — | Done. The implementation moved to `domains/elbv2` (package `elbv2`) with no API, NATS subject, storage or behaviour change. Q-30's data-plane boundary remains deferred debt; the amended ADR-0001 row makes it not a relocation prerequisite. Its cross-domain, IAM and `agents/elbv2` imports remain dependency-rule debt, with their allowlist entries now sourced from `domains/elbv2`. |
 | `handlers/iam/mock` | — | Done. Moved with the implementation to `domains/iam/mock`. |
 | `gateway` (root) | Generic HTTP route assembly, SigV4/registry authentication, the policy gate, request-audit middleware, request-context keys, and the legacy per-service dispatch switch for every service not yet registered. | Deferred: the ADR-0002 phases above (auth/IAM and request-context boundary design) gate moving the audit, context-key and legacy-dispatch mechanics. |
-| `gateway/bedrock` | — | Done. The whole package moved into `domains/ochre` in `spinifex#1266` with its package-private core and tests intact and no behaviour change. Separating the action adapters into `domains/ochre/awsapi` remains deferred, recorded as `domains/ochre` (action adapters) debt above. |
+| `gateway/bedrock` | — | Done. The whole package moved into `domains/ochre` in `spinifex#1266` with its package-private core and tests intact and no behaviour change. Separating the action adapters into `domains/ochre/awsapi` remains deferred, recorded as `domains/ochre` (action adapters) debt above. The action adapters then moved to `domains/ochre/awsapi` behind one ADR-0002 registration in `spinifex#1351`. |
 | `gateway/{iam,sts,eks,rds}` | — | Done. The action adapters moved to `domains/{iam,sts,eks,rds}/awsapi` in `spinifex#1265` with no API, NATS subject, storage or behaviour change. ADR-0001 does not make ADR-0002 registration a prerequisite for relocating a domain's adapters, as the EC2 adapter moves already showed; the root `gateway` dispatch that still imports them is recorded `gateway` (service registration) debt. |
 | `gateway/{ecs,elbv2}` | — | Done. The action adapters moved to `domains/{ecs,elbv2}/awsapi` in `spinifex#1273` with no API, NATS subject, storage or behaviour change; the root `gateway` dispatch that still imports them is recorded `gateway` (service registration) debt. |
-| `gateway/spx` | `promote_image.go` and `spx.go` dispatch (`storage.go` already consumes `contracts/operator/v1`). | Needs decision: which package owns operator-plane action dispatch registration — an `ingress/aws`-style seam for the operator plane, or `gateway` stays its home. |
-| `gateway/tagging` | — | Done. `domains/tagging` owns the read-only cross-resource `GetResources` aggregation and its API semantics (`0d830e901`, `aedd0b2c4`); AWSGW composes the owner listers, the EC2 tag index and ELBv2 (`f3545e3ea`); the root `gateway` keeps only the legacy dispatch in `gateway/tagging.go`. The package is removed, and its two allowlist rows and frozen-legacy entry are gone. Owner tag mutation capabilities remain future direction, so the RDS and ECS tag-writing edges above stay open. |
+| `gateway/spx` | — | Done. The gather helpers, storage-status config and image-promote request moved to `operator/admin` with no behaviour change and the package is removed. The `spinifex` Query dispatch and the `/admin/{method}` handlers that call them now live in `operator/admin` as one seam registration. |
+| `gateway/tagging` | — | Done. `domains/tagging` owns the read-only cross-resource `GetResources` aggregation and its API semantics (`0d830e901`, `aedd0b2c4`); AWSGW composes the owner listers, the EC2 tag index and ELBv2 (`f3545e3ea`); the dispatch moved to `domains/tagging/awsapi` and the root `gateway` no longer serves Tagging. The package is removed, and its two allowlist rows and frozen-legacy entry are gone. Owner tag mutation capabilities remain future direction, so the RDS and ECS tag-writing edges above stay open. |
 | `services/viperblockd` | — | Done. The adapter is `providers/ebs/viperblock` (with the read-only `dirty` reader), the shared vocabulary is `contracts/viperblockd/legacy/v1`, and the role is `runtime/roles/viperblockd`. `services/` is empty. |
 | `utils` | — | Done and deleted. `ec2_subject.go` went to `contracts/ec2/v1`, the vpcd protocol to `contracts/network/v1` with the typed client in `domains/network/projection`, and the master-key loader to `domains/ec2/ebs/encryption`. |
 | `vpcd` | — | Done. The IMDS local-file reader and record-space opener moved to `domains/ec2/instance/state` and are injected through `guestmetadata.LocalStateReader`/`RecordLoader` (recorded above), so `vpcd` no longer imports `daemon` directly; its inter-process boundary now exists (`contracts/network/v1`, with `domains/network/subscribers` as the consumer and `domains/network/projection` as the producer client). The package moved unchanged to `runtime/roles/vpcd` (recorded below), and the legacy directory is removed. |
@@ -696,7 +710,7 @@ ADR-0001 completion is structural, so none of these is cleared by remediating se
 | Class | Paths | Status |
 |---|---|---|
 | No missing boundary (candidate moves) | `handlers/ecs/bus` → `contracts/ecs/v1` (`a130f6b8e`); `admin/images_promote.go` → `domains/ec2/image` (`263c58167`); `admin` bootstrap account identifiers → `foundation/aws/identifiers` (`3fdaa54d9`); `admin` credential-format generators → `foundation/aws/credentials` `0368e861a`, retiring `admin` | Done. |
-| One narrow missing boundary | Done: `handlers/sts` EKS JWKS lookup → `handlers/eks` `ClusterJWKSResolver` (`254d898a3`); `gateway/tagging` → `domains/tagging` with AWSGW-composed listers (`0d830e901`, `f3545e3ea`, `aedd0b2c4`). Open: `gateway/spx` (owner of operator-plane dispatch registration) | Each needs one decision or one capability, then a mechanical move. |
+| One narrow missing boundary | Done: `handlers/sts` EKS JWKS lookup → `handlers/eks` `ClusterJWKSResolver` (`254d898a3`); `gateway/tagging` → `domains/tagging` with AWSGW-composed listers (`0d830e901`, `f3545e3ea`, `aedd0b2c4`); `gateway/spx` → `operator/admin`. Done: the `spinifex` service and `/admin` dispatch registration, owned by `operator/admin` | Each needs one decision or one capability, then a mechanical move. |
 | Broader boundary or follows another package | `gateway` root (ADR-0002 auth and request-context phase) | Recorded debt with removal conditions above. |

 Per-consumer IAM capabilities are the widest unblocker: they gate STS's IAM reads, Ochre's IAM reads, the ELBv2 role path and the ECS, EKS and RDS ensurers.
@@ -157,6 +157,7 @@ Restarting the daemon is what re-arms it: the node rebuilds its peer list from t

 > [!WARNING]
 > Do not flush the whole nftables ruleset to clear the policy. The VPC daemon writes SNAT and per-elastic-IP forward rules into the `ip nat` and `ip filter` tables and reinstalls them only when it starts, so `nft flush ruleset` breaks elastic IPs silently until the next `systemctl restart spinifex-vpcd`. Use `disable`, which only ever touches `inet spinifex_filter`.
+> On Oracle Linux with IPsec on, a flush also removes `inet spinifex_ipsec_guard`, so Geneve to a peer without an SA goes out in clear until the daemon puts the table back, within 60 seconds.

 ## The Files

@@ -51,12 +51,24 @@ sudo spx admin gpu setup
 What it does:

 1. Detects installed GPUs and their PCI IDs.
-2. Enables IOMMU in GRUB (`intel_iommu=on iommu=pt` or `amd_iommu=on iommu=pt`).
+2. Adds the IOMMU kernel arguments (`intel_iommu=on iommu=pt` or `amd_iommu=on iommu=pt`): to `GRUB_CMDLINE_LINUX_DEFAULT` with `update-grub`, or with `grubby` to every boot entry and to `/etc/kernel/cmdline`, which kernels installed later inherit.
 3. Writes the vfio udev rule (`/etc/udev/rules.d/99-spinifex-vfio.rules`).
 4. Blacklists `nouveau` (NVIDIA) and `amdgpu` (AMD) so the host kernel does not claim the device.
 5. Configures `vfio-pci` early binding in `/etc/modprobe.d/vfio-pci.conf`.
-6. Adds vfio modules to initramfs.
+6. Loads the vfio modules from the initramfs: through `/etc/initramfs-tools/modules` and `update-initramfs -u`, or through `/etc/dracut.conf.d/spinifex-vfio.conf` and `dracut -f --regenerate-all`. With dracut, each rebuilt image is checked with `lsinitrd` for `vfio_pci`, its early-load line and the `modprobe.d` files, and setup stops naming whatever is missing.

+The command finds its boot tools before changing anything, and stops naming them if the host has neither or both of `update-grub` and `grubby`, or of `update-initramfs` and `dracut`.
+
+**Oracle Linux hosts** take the `grubby` and `dracut` path: Oracle Linux keeps kernel arguments in BLS boot entries and has no `GRUB_CMDLINE_LINUX_DEFAULT`.
+`grubby --update-kernel=ALL` writes the IOMMU arguments to every entry and to `/etc/kernel/cmdline`, from which a kernel installed later builds its own entry, and dracut applies `spinifex-vfio.conf` to every initramfs it builds.
+After installing a new `kernel-uek`, check its entry before rebooting into it:
+
+```bash
+sudo grubby --info=ALL | grep -E '^(kernel|args)'
+sudo lsinitrd -k <new-kernel-version> | grep vfio
+```
+
+Whole-GPU passthrough is supported on Oracle Linux; NVIDIA MIG is not.
 If any change requires a reboot, the command prints instructions and exits:

 ```
@@ -175,7 +187,7 @@ The command signals the daemon and waits for confirmation. The GPU is released b

 ### IOMMU not active after reboot

-Verify IOMMU is enabled in BIOS/UEFI (`Intel VT-d` or `AMD-Vi`). Then check GRUB picked up the parameters:
+Verify IOMMU is enabled in BIOS/UEFI (`Intel VT-d` or `AMD-Vi`). Then check the boot loader picked up the parameters (on Oracle Linux, `sudo grubby --info=DEFAULT` lists them for the default entry):

 ```bash
 cat /proc/cmdline | grep iommu
@@ -144,6 +144,27 @@ The guest MTU moves with the setting, but a running guest holds its DHCP lease f

 **Expect launches to fail for a minute or two after re-enabling.** `RunInstances` returns `ServerInternal` with `vpc.add-nat` timing out, because the OVN flows-ready barrier cannot confirm while `ovn-controller` reprograms every tunnel for the new SAs. It clears once `ip xfrm state` settles. The error is reported as a capacity problem, which it is not.

+### Containers and nested overlays inside a guest
+
+A Docker bridge, Kubernetes pod network or other overlay inside a guest defaults to an MTU of 1500, which is larger than the guest's 1408 or 1442.
+Downloads from the internet still work for TCP on the veth and routed uplinks, because vpcd clamps the TCP MSS on the host hop (see [Performance Tuning → MTU](/docs/performance-tuning#4-mtu)).
+That was measured on a single node (env12) with a Docker container; it was not measured on a physical uplink or on a multi-node cluster.
+
+On a physical uplink there is no host hop to clamp, and UDP is never clamped, so set the inner MTU to the guest MTU.
+For Docker, in `/etc/docker/daemon.json`:
+
+```json
+{
+  "mtu": 1408,
+  "default-network-opts": {
+    "bridge": { "com.docker.network.driver.mtu": "1408" }
+  }
+}
+```
+
+Use 1442 when `ipsec_enabled = false`, and restart the Docker daemon so the default bridge is recreated.
+The gateway MTU follows `ipsec_enabled` and converges on the next reconcile pass, so change the inner MTU together with the guest MTU.
+
 ## Private vs Public Subnets

 A subnet's behavior depends on three things: whether the VPC has an Internet Gateway, whether the subnet's route table has a default route to that IGW, and whether the subnet has `MapPublicIpOnLaunch` enabled.
@@ -45,7 +45,8 @@ Each repository holds image manifests and their layers. Images are stored in obj

 - Repository lifecycle — `CreateRepository`, `DeleteRepository`, `DescribeRepositories`.
 - OCI push and pull — blobs, manifests, and tags, including cross-repository blob mounts.
- Token auth — `GetAuthorizationToken` plus bearer/basic auth on `/v2/`.
+- OCI 1.1 referrers — the referrers API, so cosign v3 signatures and other artifacts attached to an image can be listed (see [OCI referrers and image signing](#oci-referrers-and-image-signing)).
+- Token auth — `GetAuthorizationToken` plus bearer/basic auth on `/v2/`. A token is valid for 12 hours. One minted by an assumed-role session (an EKS node, for example) stays valid after that session expires and is checked against the live role on every request, so it stops working when the role is deleted or loses its ECR permissions.
 - Repository policies — set, get, and delete a repository policy document.

 **Current limitations**
@@ -176,6 +177,71 @@ Then authenticate Docker and push using the `repository_url` output, following s

 :::

+## EKS workers and multiple gateways
+
+Each gateway puts its own address in the `repositoryUri` it returns, so repositories created through different gateways can carry different hosts.
+When an EKS worker launches, the daemon that launches it writes every cluster gateway into the worker's registry mirrors, TLS trust and credential-provider config.
+A pod can then pull from any `repositoryUri` host in the cluster.
+The worker tries the gateway named in the image reference first and falls back to the others in cluster order.
+
+Limits:
+
+- The gateway list comes from the cluster config the launching daemon loaded at start.
+  A gateway added later is not trusted by existing workers; restart the launching daemon and replace the worker.
+- Peer gateways are assumed to listen on port 9999.
+  A hand-edited `awsgw` bind on a peer is not covered.
+- A gateway address that does not answer costs about 30 seconds on each cold pull while the worker's connection attempt gives up.
+  A stopped gateway on a host that is still up refuses the connection, and the worker moves to the next gateway at once.
+
+Measured on a single-node environment: pulls by `repositoryUri` and by the account's ECR DNS name, a pull rejected with `x509: certificate signed by unknown authority` when the gateway host was not in the worker config, and a 31 second pull through an unreachable first endpoint.
+Measured on a 3-node environment: a pod pulling each of the three gateways' `repositoryUri` on a worker launched with this change (two of the three failed with `x509` on a worker launched without it), and a pull of the stopped gateway's own `repositoryUri` that succeeded through another gateway within one retry (the pull itself took 147 ms after the first endpoint refused).
+
+## OCI referrers and image signing
+
+A manifest can name another manifest as its `subject`, which is how tools such as cosign v3 attach signatures and SBOMs to an image.
+The registry records the subject when the manifest is pushed and answers with an `OCI-Subject` header on the `201`.
+The `subject` does not have to exist in the repository when the referrer is pushed.
+
+`GET /v2/<name>/referrers/<digest>` lists the manifests that name `<digest>` as their subject.
+The request is authorised as `ecr:BatchGetImage` on the repository.
+The response is `200` with `Content-Type: application/vnd.oci.image.index.v1+json` and an index whose `manifests` hold each referrer's `mediaType`, `digest`, `size`, `artifactType` and annotations.
+A digest with no referrers, or one that is not stored, returns an empty index.
+A malformed digest returns `400` with `DIGEST_INVALID`.
+
+Adding `?artifactType=<type>` returns only referrers of that type and sets `OCI-Filters-Applied: artifactType`.
+The header is also set when nothing matches.
+An image manifest without an `artifactType` is listed under its config `mediaType`; an index without one has none.
+
+Deleting an image by digest (`DELETE /v2/<name>/manifests/<digest>`, `aws ecr batch-delete-image` with `imageDigest`) also deletes its referrers, and their referrers in turn.
+Deleting by tag removes only the tag and leaves the image and its referrers.
+
+cosign v3 sign and verify work against the registry, including on a child manifest of an image index copied with `cosign copy`.
+cosign v2 keeps writing its `sha256-<digest>.sig` tags and is unaffected.
+`GET` on a manifest still returns `406` when the request's `Accept` header does not include the stored `mediaType`.
+
+Limits:
+
+- `cosign copy` does not copy v3 signatures to the destination repository.
+  This is client behaviour.
+  To keep the original signature, copy with `oras cp -r <source> <mirror>`, which copies the image and its referrers in one command, or `crane copy` the signature manifest by digest after the image.
+  Otherwise, sign the copy again.
+- Referrers pushed before this version are not listed.
+  Push them again to add them to the index.
+- A crash between deleting an image and deleting its referrers can leave referrers behind that nothing lists.
+- The referrers listing is not paginated.
+- `HEAD` on the referrers route is not served.
+
+Measured on a 3-node environment with cosign v3.1.4 and v2.6.5:
+
+- cosign v3.1.4 signed a fresh image index, with one `GET /referrers` answered `200`, no `sha256-<digest>` tag written and no `406`.
+- Verify passed.
+- A copied index's child manifest signed and verified, which returned `406` before this change.
+- The listing returned the signature with its `artifactType`, the filter returned one entry for the signature's type and none for another type.
+- `aws ecr batch-delete-image` of the signed image left an empty listing and `404` for both the signature and the image.
+- cosign v2.6.5 signed and verified a fresh image.
+
+Not measured: `oras`, `notation`, a listing with many referrers, and a crash in the middle of a delete.
+
 ## Troubleshooting

 **`docker login` fails with an authorization error.** The token is account-scoped and short-lived. Re-run `aws ecr get-login-password` to mint a fresh token, and confirm the registry host in `docker login` matches your account's endpoint (`aws ecr describe-repositories` returns it).
@@ -58,8 +58,42 @@ Authentication uses **EKS access entries** (the API authentication mode) — IAM
 - **Access entries are the only auth path.** `authentication_mode` accepts `API` (the default when unset) and `API_AND_CONFIG_MAP`; bare `CONFIG_MAP` is rejected. Spinifex has no `aws-auth` ConfigMap path, so `API_AND_CONFIG_MAP` behaves as `API`: grant access with access entries.
 - **Security groups are auto-managed.** Spinifex creates the cluster and node-group security groups deterministically (e.g. `eks-cluster-<name>-nodegroup-sg`); `vpc_config.security_group_ids` is ignored. To expose a workload you add an ingress rule to the auto-managed node-group SG.
 - **The worker image is fixed.** Node groups always boot Spinifex's `eks-node` image. `ami_type` is recorded but does not select the image, and the image must be registered on the cluster before you create a node group.
+- **Workers launch with IMDSv2 required and an IMDS hop limit of 2.**
+  This matches AWS managed node groups and lets a pod without IRSA reach the node role through the SDK default credential chain.
+  A node group launch template that sets `metadata_options` overrides each field it sets, and any field it leaves unset keeps the default.
+  Existing workers keep their hop limit until they are replaced or changed with `ModifyInstanceMetadataOptions`.
 - **Default Kubernetes version is `1.32`.**

+### Control plane readiness and kubectl authentication
+
+Each control-plane VM authenticates `kubectl` tokens through a webhook that calls the AWS gateway on the host the VM runs on.
+When that gateway is stopped, the webhook cannot validate tokens and `kubectl` reports `Unauthorized` for the share of requests that the load balancer sends to that control plane.
+
+The webhook serves `/readyz` on port 8444 of each control-plane VM.
+It returns 200 only when the host gateway answers a TLS handshake and the local API server accepts a TCP connection, and 503 otherwise.
+The API server target group health-checks `HTTP :8444/readyz` every 5 seconds with a healthy and unhealthy threshold of 2.
+The target group still forwards TCP on 6443.
+A control plane that cannot authenticate leaves rotation, and returns when its gateway is back.
+
+Measured on a three-node cluster, one control plane per host, with `kubectl get ns` run continuously after stopping one gateway:
+
+| Gateway stopped | Result |
+| --- | --- |
+| Host without the cluster load balancer VM, first run | 4 of 20 calls failed in the first round, 0 of 20 in the next round 54 s after the stop |
+| Host without the cluster load balancer VM, second run | 2 failures in 100 s, at 3 s and 8 s after the stop |
+| Host running the cluster load balancer VM | Not fixed: 20 of 60 calls failed over 100 s |
+
+Before this change the same test failed 3, 4 and 3 calls of 10 per stopped gateway.
+
+Limits:
+
+- Only clusters created with the current `eks-node` image get the readiness listener, and only clusters created from now on get the health check on the target group.
+  Older clusters keep the TCP check on 6443 and are not changed.
+- A cluster created with an older image gets the health check on its target group but no listener on 8444, so every control plane fails the check and the load balancer fails open to all of them.
+- A stopped gateway on the host that runs the cluster load balancer VM is not covered.
+  The load balancer VM reports target health through its own host's gateway, so while that gateway is down the unhealthy control plane stays in rotation.
+- Not measured: more than one control plane per host, a physical-node failure instead of a stopped gateway, and the target health state as `DescribeTargetHealth` reports it, because the load balancer and control-plane VMs belong to the system account.
+
 ## Prerequisites

 > [!IMPORTANT]
@@ -123,7 +157,7 @@ aws eks describe-cluster --name demo --query 'cluster.status'
 aws eks create-nodegroup \
  --cluster-name demo \
  --nodegroup-name default \
-  --node-role-arn arn:aws:iam::000000000000:role/eks-node-role \
+  --node-role arn:aws:iam::000000000000:role/eks-node-role \
  --subnets subnet-aaaa subnet-bbbb \
  --scaling-config minSize=1,maxSize=2,desiredSize=1 \
  --instance-types t3.medium
@@ -256,6 +290,10 @@ The workbooks form a ladder. **eks-quickstart** is a NodePort demo; **eks-https-

 **`kubectl` cannot connect.** Re-run `aws eks update-kubeconfig --name <name>` to refresh the endpoint and CA, and confirm the cluster is `ACTIVE`. `aws sts get-caller-identity` verifies your credentials reach the Spinifex endpoint.

+**`kubectl` returns `Unauthorized` for some requests and works for others.**
+One control plane cannot reach its host's AWS gateway, so check `systemctl status spinifex-awsgw` on each node.
+Control plane readiness and kubectl authentication above lists which clusters recover on their own.
+
 **Workers stay `NotReady`.** Almost always an egress or IAM problem:
 - Confirm the workers can pull images — a default route to an Internet Gateway (public) or NAT Gateway (private).
 - Confirm the node role has `AmazonEKSWorkerNodePolicy`, `AmazonEKS_CNI_Policy`, and `AmazonEC2ContainerRegistryReadOnly`.
@@ -17,7 +17,7 @@ tags:

 ## Overview

-The platform serves **392 operations** across the AWS APIs below. Every page names the operations its service implements from the pinned model, generated from the dispatch tables on each build rather than written by hand.
+The platform serves **405 operations** across the AWS APIs below. Every page names the operations its service implements from the pinned model, generated from the dispatch tables on each build rather than written by hand.

 | Service | Operations |
 |---|---:|
@@ -29,6 +29,7 @@ The platform serves **392 operations** across the AWS APIs below. Every page nam
 | [ELBv2](/coverage/elbv2) | 36 |
 | [IAM](/coverage/iam) | 81 |
 | [RDS](/coverage/rds) | 27 |
+| [Route 53](/coverage/route53) | 13 |
 | [S3](/coverage/s3) | 26 |
 | [STS](/coverage/sts) | 5 |
-| **Total** | **392** |
+| **Total** | **405** |
@@ -2,8 +2,8 @@
  "schema_version": 1,
  "model_source": {
    "repository": "github.com/aws/api-models-aws",
-    "commit": "9b1a4929274443d8671ea8acfda7afbe856e29e9",
-    "commit_date": "2026-09-25"
+    "commit": "e10d6a1c5ac36af7c0c61d1f67b3c3ebbe41de49",
+    "commit_date": "2026-10-28"
  },
  "services": [
    {
@@ -300,6 +300,7 @@
          "DeleteCapacityManagerDataExport",
          "DeleteCarrierGateway",
          "DeleteClientVpnEndpoint",
+          "DeleteClientVpnEndpointAuthorizationPolicy",
          "DeleteClientVpnRoute",
          "DeleteCoipCidr",
          "DeleteCoipPool",
@@ -681,7 +682,9 @@
          "GetCapacityManagerMetricData",
          "GetCapacityManagerMetricDimensions",
          "GetCapacityManagerMonitoredTagKeys",
+          "GetCapacityManagerSupportabilityScores",
          "GetCapacityReservationUsage",
+          "GetClientVpnEndpointAuthorizationPolicy",
          "GetCoipPoolUsage",
          "GetConsoleOutput",
          "GetConsoleScreenshot",
@@ -695,6 +698,7 @@
          "GetHostReservationPurchasePreview",
          "GetImageAncestry",
          "GetImageBlockPublicAccessState",
+          "GetImageFindings",
          "GetInstanceMetadataDefaults",
          "GetInstanceTpmEkPub",
          "GetInstanceTypesFromInstanceRequirements",
@@ -734,6 +738,7 @@
          "GetSnapshotBlockPublicAccessState",
          "GetSpotPlacementScores",
          "GetSubnetCidrReservations",
+          "GetTransitGatewayAttachmentNatConfiguration",
          "GetTransitGatewayAttachmentPropagations",
          "GetTransitGatewayMeteringPolicyEntries",
          "GetTransitGatewayMulticastDomainAssociations",
@@ -766,6 +771,7 @@
          "ModifyCapacityReservation",
          "ModifyCapacityReservationFleet",
          "ModifyClientVpnEndpoint",
+          "ModifyClientVpnEndpointAuthorizationPolicy",
          "ModifyDefaultCreditSpecification",
          "ModifyEbsDefaultKmsKeyId",
          "ModifyFleet",
@@ -814,6 +820,7 @@
          "ModifyTrafficMirrorFilterRule",
          "ModifyTrafficMirrorSession",
          "ModifyTransitGateway",
+          "ModifyTransitGatewayAttachmentNatConfiguration",
          "ModifyTransitGatewayMeteringPolicy",
          "ModifyTransitGatewayPolicyTableEntry",
          "ModifyTransitGatewayPrefixListReference",
@@ -1335,6 +1342,7 @@
          "DeleteCapacityManagerDataExport",
          "DeleteCarrierGateway",
          "DeleteClientVpnEndpoint",
+          "DeleteClientVpnEndpointAuthorizationPolicy",
          "DeleteClientVpnRoute",
          "DeleteCoipCidr",
          "DeleteCoipPool",
@@ -1653,7 +1661,9 @@
          "GetCapacityManagerMetricData",
          "GetCapacityManagerMetricDimensions",
          "GetCapacityManagerMonitoredTagKeys",
+          "GetCapacityManagerSupportabilityScores",
          "GetCapacityReservationUsage",
+          "GetClientVpnEndpointAuthorizationPolicy",
          "GetCoipPoolUsage",
          "GetConsoleScreenshot",
          "GetDeclarativePoliciesReportSummary",
@@ -1665,6 +1675,7 @@
          "GetHostReservationPurchasePreview",
          "GetImageAncestry",
          "GetImageBlockPublicAccessState",
+          "GetImageFindings",
          "GetInstanceMetadataDefaults",
          "GetInstanceTpmEkPub",
          "GetInstanceTypesFromInstanceRequirements",
@@ -1701,6 +1712,7 @@
          "GetSnapshotBlockPublicAccessState",
          "GetSpotPlacementScores",
          "GetSubnetCidrReservations",
+          "GetTransitGatewayAttachmentNatConfiguration",
          "GetTransitGatewayAttachmentPropagations",
          "GetTransitGatewayMeteringPolicyEntries",
          "GetTransitGatewayMulticastDomainAssociations",
@@ -1732,6 +1744,7 @@
          "ModifyCapacityReservation",
          "ModifyCapacityReservationFleet",
          "ModifyClientVpnEndpoint",
+          "ModifyClientVpnEndpointAuthorizationPolicy",
          "ModifyDefaultCreditSpecification",
          "ModifyEbsDefaultKmsKeyId",
          "ModifyFleet",
@@ -1773,6 +1786,7 @@
          "ModifyTrafficMirrorFilterRule",
          "ModifyTrafficMirrorSession",
          "ModifyTransitGateway",
+          "ModifyTransitGatewayAttachmentNatConfiguration",
          "ModifyTransitGatewayMeteringPolicy",
          "ModifyTransitGatewayPolicyTableEntry",
          "ModifyTransitGatewayPrefixListReference",
@@ -1897,6 +1911,7 @@
          "DescribeImages",
          "DescribePullThroughCacheRules",
          "DescribeRegistry",
+          "DescribeReplicationBackfillResults",
          "DescribeRepositories",
          "DescribeRepositoryCreationTemplates",
          "GetAccountSetting",
@@ -1912,6 +1927,7 @@
          "ListImageReferrers",
          "ListImages",
          "ListPullTimeUpdateExclusions",
+          "ListReplicationBackfills",
          "ListTagsForResource",
          "PutAccountSetting",
          "PutImage",
@@ -1926,6 +1942,8 @@
          "SetRepositoryPolicy",
          "StartImageScan",
          "StartLifecyclePolicyPreview",
+          "StartReplicationBackfill",
+          "StopReplicationBackfill",
          "TagResource",
          "UntagResource",
          "UpdateImageStorageClass",
@@ -2031,14 +2049,18 @@
          "DescribeImageReplicationStatus",
          "DescribeImageSigningStatus",
          "DescribePullThroughCacheRules",
+          "DescribeReplicationBackfillResults",
          "DescribeRepositoryCreationTemplates",
          "GetAccountSetting",
          "GetSigningConfiguration",
          "ListImageReferrers",
          "ListPullTimeUpdateExclusions",
+          "ListReplicationBackfills",
          "PutAccountSetting",
          "PutSigningConfiguration",
          "RegisterPullTimeUpdateExclusion",
+          "StartReplicationBackfill",
+          "StopReplicationBackfill",
          "UpdateImageStorageClass",
          "UpdatePullThroughCacheRule",
          "UpdateRepositoryCreationTemplate",
@@ -2058,11 +2080,13 @@
      "note": "",
      "operations": {
        "modelled": [
+          "AssociateGatewayWithCluster",
          "ContinueServiceDeployment",
          "CreateCapacityProvider",
          "CreateCluster",
          "CreateDaemon",
          "CreateExpressGatewayService",
+          "CreateGateway",
          "CreateService",
          "CreateTaskSet",
          "DeleteAccountSetting",
@@ -2072,6 +2096,7 @@
          "DeleteDaemon",
          "DeleteDaemonTaskDefinition",
          "DeleteExpressGatewayService",
+          "DeleteGateway",
          "DeleteService",
          "DeleteTaskDefinitions",
          "DeleteTaskSet",
@@ -2085,12 +2110,16 @@
          "DescribeDaemonRevisions",
          "DescribeDaemonTaskDefinition",
          "DescribeExpressGatewayService",
+          "DescribeGateway",
+          "DescribeGatewayResources",
+          "DescribeGatewayRoute",
          "DescribeServiceDeployments",
          "DescribeServiceRevisions",
          "DescribeServices",
          "DescribeTaskDefinition",
          "DescribeTaskSets",
          "DescribeTasks",
+          "DisassociateGatewayFromCluster",
          "DiscoverPollEndpoint",
          "ExecuteCommand",
          "GetTaskProtection",
@@ -2101,6 +2130,9 @@
          "ListDaemonDeployments",
          "ListDaemonTaskDefinitions",
          "ListDaemons",
+          "ListGatewayClusterAssociations",
+          "ListGatewayRoutes",
+          "ListGateways",
          "ListServiceDeployments",
          "ListServices",
          "ListServicesByNamespace",
@@ -2131,6 +2163,7 @@
          "UpdateContainerInstancesState",
          "UpdateDaemon",
          "UpdateExpressGatewayService",
+          "UpdateGateway",
          "UpdateService",
          "UpdateServicePrimaryTaskSet",
          "UpdateTaskProtection",
@@ -2219,15 +2252,18 @@
        ],
        "unsupported": [],
        "missing": [
+          "AssociateGatewayWithCluster",
          "ContinueServiceDeployment",
          "CreateDaemon",
          "CreateExpressGatewayService",
+          "CreateGateway",
          "CreateTaskSet",
          "DeleteAccountSetting",
          "DeleteAttributes",
          "DeleteDaemon",
          "DeleteDaemonTaskDefinition",
          "DeleteExpressGatewayService",
+          "DeleteGateway",
          "DeleteTaskDefinitions",
          "DeleteTaskSet",
          "DescribeDaemon",
@@ -2235,9 +2271,13 @@
          "DescribeDaemonRevisions",
          "DescribeDaemonTaskDefinition",
          "DescribeExpressGatewayService",
+          "DescribeGateway",
+          "DescribeGatewayResources",
+          "DescribeGatewayRoute",
          "DescribeServiceDeployments",
          "DescribeServiceRevisions",
          "DescribeTaskSets",
+          "DisassociateGatewayFromCluster",
          "DiscoverPollEndpoint",
          "ExecuteCommand",
          "GetTaskProtection",
@@ -2245,6 +2285,9 @@
          "ListDaemonDeployments",
          "ListDaemonTaskDefinitions",
          "ListDaemons",
+          "ListGatewayClusterAssociations",
+          "ListGatewayRoutes",
+          "ListGateways",
          "ListServiceDeployments",
          "PutAccountSettingDefault",
          "PutAttributes",
@@ -2257,6 +2300,7 @@
          "UpdateContainerAgent",
          "UpdateDaemon",
          "UpdateExpressGatewayService",
+          "UpdateGateway",
          "UpdateServicePrimaryTaskSet",
          "UpdateTaskProtection",
          "UpdateTaskSet"
@@ -3099,6 +3143,7 @@
          "ApplyPendingMaintenanceAction",
          "AuthorizeDBSecurityGroupIngress",
          "BacktrackDBCluster",
+          "CancelDBCapacityReservation",
          "CancelExportTask",
          "CopyDBClusterParameterGroup",
          "CopyDBClusterSnapshot",
@@ -3107,6 +3152,7 @@
          "CopyOptionGroup",
          "CreateBlueGreenDeployment",
          "CreateCustomDBEngineVersion",
+          "CreateDBCapacityReservation",
          "CreateDBCluster",
          "CreateDBClusterEndpoint",
          "CreateDBClusterParameterGroup",
@@ -3150,6 +3196,7 @@
          "DescribeAccountAttributes",
          "DescribeBlueGreenDeployments",
          "DescribeCertificates",
+          "DescribeDBCapacityReservations",
          "DescribeDBClusterAutomatedBackups",
          "DescribeDBClusterBacktracks",
          "DescribeDBClusterEndpoints",
@@ -3204,6 +3251,7 @@
          "ModifyCertificates",
          "ModifyCurrentDBClusterCapacity",
          "ModifyCustomDBEngineVersion",
+          "ModifyDBCapacityReservation",
          "ModifyDBCluster",
          "ModifyDBClusterEndpoint",
          "ModifyDBClusterParameterGroup",
@@ -3354,6 +3402,7 @@
          "ApplyPendingMaintenanceAction",
          "AuthorizeDBSecurityGroupIngress",
          "BacktrackDBCluster",
+          "CancelDBCapacityReservation",
          "CancelExportTask",
          "CopyDBClusterParameterGroup",
          "CopyDBClusterSnapshot",
@@ -3362,6 +3411,7 @@
          "CopyOptionGroup",
          "CreateBlueGreenDeployment",
          "CreateCustomDBEngineVersion",
+          "CreateDBCapacityReservation",
          "CreateDBClusterEndpoint",
          "CreateDBClusterParameterGroup",
          "CreateDBClusterSnapshot",
@@ -3392,6 +3442,7 @@
          "DescribeAccountAttributes",
          "DescribeBlueGreenDeployments",
          "DescribeCertificates",
+          "DescribeDBCapacityReservations",
          "DescribeDBClusterAutomatedBackups",
          "DescribeDBClusterBacktracks",
          "DescribeDBClusterEndpoints",
@@ -3433,6 +3484,7 @@
          "ModifyCertificates",
          "ModifyCurrentDBClusterCapacity",
          "ModifyCustomDBEngineVersion",
+          "ModifyDBCapacityReservation",
          "ModifyDBClusterEndpoint",
          "ModifyDBClusterParameterGroup",
          "ModifyDBClusterSnapshotAttribute",
@@ -3483,6 +3535,180 @@
        ]
      }
    },
+    {
+      "service": "route53",
+      "api_version": "2013-04-01",
+      "opaque": false,
+      "note": "",
+      "operations": {
+        "modelled": [
+          "ActivateKeySigningKey",
+          "AssociateVPCWithHostedZone",
+          "ChangeCidrCollection",
+          "ChangeResourceRecordSets",
+          "ChangeTagsForResource",
+          "CreateCidrCollection",
+          "CreateHealthCheck",
+          "CreateHostedZone",
+          "CreateKeySigningKey",
+          "CreateQueryLoggingConfig",
+          "CreateReusableDelegationSet",
+          "CreateTrafficPolicy",
+          "CreateTrafficPolicyInstance",
+          "CreateTrafficPolicyVersion",
+          "CreateVPCAssociationAuthorization",
+          "DeactivateKeySigningKey",
+          "DeleteCidrCollection",
+          "DeleteHealthCheck",
+          "DeleteHostedZone",
+          "DeleteKeySigningKey",
+          "DeleteQueryLoggingConfig",
+          "DeleteReusableDelegationSet",
+          "DeleteTrafficPolicy",
+          "DeleteTrafficPolicyInstance",
+          "DeleteVPCAssociationAuthorization",
+          "DisableHostedZoneDNSSEC",
+          "DisassociateVPCFromHostedZone",
+          "EnableHostedZoneDNSSEC",
+          "GetAccountLimit",
+          "GetChange",
+          "GetCheckerIpRanges",
+          "GetDNSSEC",
+          "GetGeoLocation",
+          "GetHealthCheck",
+          "GetHealthCheckCount",
+          "GetHealthCheckLastFailureReason",
+          "GetHealthCheckStatus",
+          "GetHostedZone",
+          "GetHostedZoneCount",
+          "GetHostedZoneLimit",
+          "GetQueryLoggingConfig",
+          "GetReusableDelegationSet",
+          "GetReusableDelegationSetLimit",
+          "GetTrafficPolicy",
+          "GetTrafficPolicyInstance",
+          "GetTrafficPolicyInstanceCount",
+          "ListCidrBlocks",
+          "ListCidrCollections",
+          "ListCidrLocations",
+          "ListGeoLocations",
+          "ListHealthChecks",
+          "ListHostedZones",
+          "ListHostedZonesByName",
+          "ListHostedZonesByVPC",
+          "ListQueryLoggingConfigs",
+          "ListResourceRecordSets",
+          "ListReusableDelegationSets",
+          "ListTagsForResource",
+          "ListTagsForResources",
+          "ListTrafficPolicies",
+          "ListTrafficPolicyInstances",
+          "ListTrafficPolicyInstancesByHostedZone",
+          "ListTrafficPolicyInstancesByPolicy",
+          "ListTrafficPolicyVersions",
+          "ListVPCAssociationAuthorizations",
+          "TestDNSAnswer",
+          "UpdateHealthCheck",
+          "UpdateHostedZoneComment",
+          "UpdateHostedZoneFeatures",
+          "UpdateTrafficPolicyComment",
+          "UpdateTrafficPolicyInstance"
+        ],
+        "registered": [
+          "ChangeResourceRecordSets",
+          "ChangeTagsForResource",
+          "CreateHostedZone",
+          "DeleteHostedZone",
+          "GetChange",
+          "GetHostedZone",
+          "GetHostedZoneCount",
+          "ListHostedZones",
+          "ListHostedZonesByName",
+          "ListResourceRecordSets",
+          "ListTagsForResource",
+          "ListTagsForResources",
+          "UpdateHostedZoneComment"
+        ],
+        "implemented": [
+          "ChangeResourceRecordSets",
+          "ChangeTagsForResource",
+          "CreateHostedZone",
+          "DeleteHostedZone",
+          "GetChange",
+          "GetHostedZone",
+          "GetHostedZoneCount",
+          "ListHostedZones",
+          "ListHostedZonesByName",
+          "ListResourceRecordSets",
+          "ListTagsForResource",
+          "ListTagsForResources",
+          "UpdateHostedZoneComment"
+        ],
+        "stubbed": [],
+        "unsupported": [],
+        "missing": [
+          "ActivateKeySigningKey",
+          "AssociateVPCWithHostedZone",
+          "ChangeCidrCollection",
+          "CreateCidrCollection",
+          "CreateHealthCheck",
+          "CreateKeySigningKey",
+          "CreateQueryLoggingConfig",
+          "CreateReusableDelegationSet",
+          "CreateTrafficPolicy",
+          "CreateTrafficPolicyInstance",
+          "CreateTrafficPolicyVersion",
+          "CreateVPCAssociationAuthorization",
+          "DeactivateKeySigningKey",
+          "DeleteCidrCollection",
+          "DeleteHealthCheck",
+          "DeleteKeySigningKey",
+          "DeleteQueryLoggingConfig",
+          "DeleteReusableDelegationSet",
+          "DeleteTrafficPolicy",
+          "DeleteTrafficPolicyInstance",
+          "DeleteVPCAssociationAuthorization",
+          "DisableHostedZoneDNSSEC",
+          "DisassociateVPCFromHostedZone",
+          "EnableHostedZoneDNSSEC",
+          "GetAccountLimit",
+          "GetCheckerIpRanges",
+          "GetDNSSEC",
+          "GetGeoLocation",
+          "GetHealthCheck",
+          "GetHealthCheckCount",
+          "GetHealthCheckLastFailureReason",
+          "GetHealthCheckStatus",
+          "GetHostedZoneLimit",
+          "GetQueryLoggingConfig",
+          "GetReusableDelegationSet",
+          "GetReusableDelegationSetLimit",
+          "GetTrafficPolicy",
+          "GetTrafficPolicyInstance",
+          "GetTrafficPolicyInstanceCount",
+          "ListCidrBlocks",
+          "ListCidrCollections",
+          "ListCidrLocations",
+          "ListGeoLocations",
+          "ListHealthChecks",
+          "ListHostedZonesByVPC",
+          "ListQueryLoggingConfigs",
+          "ListReusableDelegationSets",
+          "ListTrafficPolicies",
+          "ListTrafficPolicyInstances",
+          "ListTrafficPolicyInstancesByHostedZone",
+          "ListTrafficPolicyInstancesByPolicy",
+          "ListTrafficPolicyVersions",
+          "ListVPCAssociationAuthorizations",
+          "TestDNSAnswer",
+          "UpdateHealthCheck",
+          "UpdateHostedZoneFeatures",
+          "UpdateTrafficPolicyComment",
+          "UpdateTrafficPolicyInstance"
+        ],
+        "extra": []
+      }
+    },
    {
      "service": "s3",
      "api_version": "2006-03-01",
@@ -28,6 +28,21 @@ Spot Instance Requests are a mock over the on-demand `RunInstances` path. A requ

 An AMI cannot be shared with another account or made public. `ModifyImageAttribute` refuses `LaunchPermission`, `UserIds`, `UserGroups`, `OrganizationArns`, `OrganizationalUnitArns` and `OperationType`, and `ResetImageAttribute` resets only `description`, not `launchPermission`. Each refusal is `InvalidParameterValue`. Terraform's `aws_ami_launch_permission` therefore cannot be created. `description` is the only image attribute that can be modified or reset.

+### RegisterImage Registers From a Snapshot
+
+`RegisterImage` takes a root EBS snapshot in `BlockDeviceMappings` and writes an image that points at it.
+`EnaSupport` is supported, with `true` and `false` both accepted and stored on the image.
+`DescribeImages` returns `EnaSupport` when the image has one and leaves the member out when it was never set, as AWS does.
+`CopyImage` carries the value to the copy, and `CreateImage` sets it to `true` because every instance reports ENA support.
+`RunInstances` does not read the flag, so an image registered with `EnaSupport=false` is not held back from instance types that need ENA on AWS, since guests use virtio networking either way.
+
+`RegisterImage` refuses `ImageLocation`, `KernelId`, `RamdiskId`, `TpmSupport`, `ImdsSupport` and `SriovNetSupport`, and any `VirtualizationType` other than `hvm`.
+Each refusal is `InvalidParameterValue`, and the message names the parameter and the value, for example `Value (aki-123) for parameter KernelId is invalid. KernelId is not supported.`
+The same form is used for a bad `Architecture`, `BootMode` or `Name` length.
+A missing `Name`, `BlockDeviceMappings` or root snapshot answers `MissingParameter` and names what is missing.
+Registering with `--ena-support`, `--no-ena-support` and neither, copying the image, and three refusals were run on a three-node cluster with the AWS CLI.
+Launching an instance from an `EnaSupport` image was not run, so the launch path is unchanged by construction and not by test.
+
 ### Operations

 | Operation |
@@ -5,3 +5,18 @@ Spot Instance Requests are a mock over the on-demand `RunInstances` path. A requ
 ### AMI Launch Permissions Are Refused

 An AMI cannot be shared with another account or made public. `ModifyImageAttribute` refuses `LaunchPermission`, `UserIds`, `UserGroups`, `OrganizationArns`, `OrganizationalUnitArns` and `OperationType`, and `ResetImageAttribute` resets only `description`, not `launchPermission`. Each refusal is `InvalidParameterValue`. Terraform's `aws_ami_launch_permission` therefore cannot be created. `description` is the only image attribute that can be modified or reset.
+
+### RegisterImage Registers From a Snapshot
+
+`RegisterImage` takes a root EBS snapshot in `BlockDeviceMappings` and writes an image that points at it.
+`EnaSupport` is supported, with `true` and `false` both accepted and stored on the image.
+`DescribeImages` returns `EnaSupport` when the image has one and leaves the member out when it was never set, as AWS does.
+`CopyImage` carries the value to the copy, and `CreateImage` sets it to `true` because every instance reports ENA support.
+`RunInstances` does not read the flag, so an image registered with `EnaSupport=false` is not held back from instance types that need ENA on AWS, since guests use virtio networking either way.
+
+`RegisterImage` refuses `ImageLocation`, `KernelId`, `RamdiskId`, `TpmSupport`, `ImdsSupport` and `SriovNetSupport`, and any `VirtualizationType` other than `hvm`.
+Each refusal is `InvalidParameterValue`, and the message names the parameter and the value, for example `Value (aki-123) for parameter KernelId is invalid. KernelId is not supported.`
+The same form is used for a bad `Architecture`, `BootMode` or `Name` length.
+A missing `Name`, `BlockDeviceMappings` or root snapshot answers `MissingParameter` and names what is missing.
+Registering with `--ena-support`, `--no-ena-support` and neither, copying the image, and three refusals were run on a three-node cluster with the AWS CLI.
+Launching an instance from an `EnaSupport` image was not run, so the launch path is unchanged by construction and not by test.
@@ -24,6 +24,74 @@ Spinifex implements **36 operations** in the ELBv2 `2015-12-01` API model.

 The data plane is a system-managed load balancer VM, launched automatically during `CreateLoadBalancer`. Application Load Balancers run HAProxy for the L7 surface — rules, fixed responses and redirects over HTTP and HTTPS. Network Load Balancers run nginx `stream` for the L4 surface — TCP, UDP, TLS and TCP_UDP — because HAProxy cannot load-balance UDP. The agent selects the engine from the configuration the control plane delivers, so the choice follows the load balancer type and is not separately configurable.

+### Target health and routing
+
+Traffic goes only to targets that are healthy or still `initial`.
+A target that has failed its health check is removed from the data plane on the next render and put back when it passes again.
+When every target in a group is unhealthy, the load balancer fails open and sends traffic to all of them rather than refusing it.
+A Network Load Balancer renders every registered server in that case, and an Application Load Balancer falls back to unchecked backup servers that HAProxy uses only when every checked server is down.
+`DescribeTargetHealth` keeps reporting the targets as unhealthy while this happens.
+Draining targets and targets without an address never receive traffic.
+
+An `initial` target receives traffic before its first successful check.
+This differs from AWS, where a target takes no traffic until it passes.
+Health state is reset to `initial` for every target when the control plane restarts, so excluding `initial` targets would drop all traffic after a restart.
+
+Changing the health check of a target group with `ModifyTargetGroup` takes effect on the next render of every load balancer that uses it.
+The time from a backend failing to traffic stopping is the health check interval times the unhealthy threshold, plus the agent poll of about 5 seconds and the heartbeat that reports the result.
+Recovery takes the interval times the healthy threshold, plus the same delays.
+
+Measured on a three-node cluster with a 5 second interval and thresholds of 2:
+
+| Event | Time |
+| --- | --- |
+| Backend starts failing to reported `unhealthy` | 21 s |
+| Backend recovers to reported `healthy` | 15 s, serving again 6 s later |
+| Every backend failing to reported `unhealthy` | 12 s, fail open immediate |
+
+A TCP connection held through a Network Load Balancer survived four configuration reloads while the other backend changed state, because the agent applies nginx with a graceful reload and HAProxy with a graceful restart.
+The HAProxy configuration with the backup servers was accepted by a live Application Load Balancer, but its rendered text was not dumped and checked by hand.
+
+### Forwarding headers
+
+An Application Load Balancer sets three headers on every HTTP and HTTPS request it forwards.
+`X-Forwarded-For` carries the client address, `X-Forwarded-Proto` is `http` or `https` for the listener that accepted the request, and `X-Forwarded-Port` is that listener's port.
+`X-Forwarded-Proto` and `X-Forwarded-Port` are always overwritten, so a value sent by the client never reaches the target.
+Network Load Balancers do not touch the stream.
+
+`routing.http.xff_header_processing.mode` decides what happens to `X-Forwarded-For`:
+
+| Mode | Behaviour |
+| --- | --- |
+| `append` (default) | Keeps what the client sent and adds the client address last, as one comma-separated line such as `203.0.113.9, 192.0.2.10`. |
+| `preserve` | Forwards the client's value unchanged and adds nothing. The header is absent when the client sent none. |
+| `remove` | Drops the header, including any client value. |
+
+`routing.http.xff_client_port.enabled` (default `false`) appends the client's source port in `append` mode, as `192.0.2.10:54321`.
+An IPv6 client is written `[2001:db8::1]:54321`.
+The setting has no effect in `preserve` and `remove` modes.
+
+`ModifyLoadBalancerAttributes` accepts only `append`, `preserve` or `remove` for the mode and `true` or `false` for the client port, and rejects anything else with `ValidationError`.
+A change re-renders the load balancer configuration, and a call that leaves both values as they were does not.
+In a live test each change reached the target within one 3 second poll.
+
+A client that sends `X-Forwarded-For` more than once has its lines joined into the single comma-separated value.
+At most eight client lines are joined, and any beyond the eighth are dropped, while the client address is still added last.
+
+Every other `routing.http.*` attribute is stored and returned by `DescribeLoadBalancerAttributes` but has no effect on the data plane.
+
+Measured on a three-node cluster, with the caller on the first node:
+
+| Listener | `X-Forwarded-For` the target saw |
+| --- | --- |
+| Internet-facing HTTP | The node's real address, `192.168.0.120` |
+| Internet-facing HTTPS | The same address, with `https` and port 443 |
+| Internal HTTP | The calling instance's private address |
+| Internet-facing HTTP with a client value of `203.0.113.9` | `203.0.113.9, 192.168.0.120` on one line |
+
+A packet capture on the load balancer's interface showed HAProxy receiving the real client address, so nothing translates it first.
+Behaviour against AWS was not compared, and the cases above were not run with an IPv6 client.
+
 ### Operations

 | Operation |
@@ -1,3 +1,71 @@
 ### Two data planes

 The data plane is a system-managed load balancer VM, launched automatically during `CreateLoadBalancer`. Application Load Balancers run HAProxy for the L7 surface — rules, fixed responses and redirects over HTTP and HTTPS. Network Load Balancers run nginx `stream` for the L4 surface — TCP, UDP, TLS and TCP_UDP — because HAProxy cannot load-balance UDP. The agent selects the engine from the configuration the control plane delivers, so the choice follows the load balancer type and is not separately configurable.
+
+### Target health and routing
+
+Traffic goes only to targets that are healthy or still `initial`.
+A target that has failed its health check is removed from the data plane on the next render and put back when it passes again.
+When every target in a group is unhealthy, the load balancer fails open and sends traffic to all of them rather than refusing it.
+A Network Load Balancer renders every registered server in that case, and an Application Load Balancer falls back to unchecked backup servers that HAProxy uses only when every checked server is down.
+`DescribeTargetHealth` keeps reporting the targets as unhealthy while this happens.
+Draining targets and targets without an address never receive traffic.
+
+An `initial` target receives traffic before its first successful check.
+This differs from AWS, where a target takes no traffic until it passes.
+Health state is reset to `initial` for every target when the control plane restarts, so excluding `initial` targets would drop all traffic after a restart.
+
+Changing the health check of a target group with `ModifyTargetGroup` takes effect on the next render of every load balancer that uses it.
+The time from a backend failing to traffic stopping is the health check interval times the unhealthy threshold, plus the agent poll of about 5 seconds and the heartbeat that reports the result.
+Recovery takes the interval times the healthy threshold, plus the same delays.
+
+Measured on a three-node cluster with a 5 second interval and thresholds of 2:
+
+| Event | Time |
+| --- | --- |
+| Backend starts failing to reported `unhealthy` | 21 s |
+| Backend recovers to reported `healthy` | 15 s, serving again 6 s later |
+| Every backend failing to reported `unhealthy` | 12 s, fail open immediate |
+
+A TCP connection held through a Network Load Balancer survived four configuration reloads while the other backend changed state, because the agent applies nginx with a graceful reload and HAProxy with a graceful restart.
+The HAProxy configuration with the backup servers was accepted by a live Application Load Balancer, but its rendered text was not dumped and checked by hand.
+
+### Forwarding headers
+
+An Application Load Balancer sets three headers on every HTTP and HTTPS request it forwards.
+`X-Forwarded-For` carries the client address, `X-Forwarded-Proto` is `http` or `https` for the listener that accepted the request, and `X-Forwarded-Port` is that listener's port.
+`X-Forwarded-Proto` and `X-Forwarded-Port` are always overwritten, so a value sent by the client never reaches the target.
+Network Load Balancers do not touch the stream.
+
+`routing.http.xff_header_processing.mode` decides what happens to `X-Forwarded-For`:
+
+| Mode | Behaviour |
+| --- | --- |
+| `append` (default) | Keeps what the client sent and adds the client address last, as one comma-separated line such as `203.0.113.9, 192.0.2.10`. |
+| `preserve` | Forwards the client's value unchanged and adds nothing. The header is absent when the client sent none. |
+| `remove` | Drops the header, including any client value. |
+
+`routing.http.xff_client_port.enabled` (default `false`) appends the client's source port in `append` mode, as `192.0.2.10:54321`.
+An IPv6 client is written `[2001:db8::1]:54321`.
+The setting has no effect in `preserve` and `remove` modes.
+
+`ModifyLoadBalancerAttributes` accepts only `append`, `preserve` or `remove` for the mode and `true` or `false` for the client port, and rejects anything else with `ValidationError`.
+A change re-renders the load balancer configuration, and a call that leaves both values as they were does not.
+In a live test each change reached the target within one 3 second poll.
+
+A client that sends `X-Forwarded-For` more than once has its lines joined into the single comma-separated value.
+At most eight client lines are joined, and any beyond the eighth are dropped, while the client address is still added last.
+
+Every other `routing.http.*` attribute is stored and returned by `DescribeLoadBalancerAttributes` but has no effect on the data plane.
+
+Measured on a three-node cluster, with the caller on the first node:
+
+| Listener | `X-Forwarded-For` the target saw |
+| --- | --- |
+| Internet-facing HTTP | The node's real address, `192.168.0.120` |
+| Internet-facing HTTPS | The same address, with `https` and port 443 |
+| Internal HTTP | The calling instance's private address |
+| Internet-facing HTTP with a client value of `203.0.113.9` | `203.0.113.9, 192.168.0.120` on one line |
+
+A packet capture on the load balancer's interface showed HAProxy receiving the real client address, so nothing translates it first.
+Behaviour against AWS was not compared, and the cases above were not run with an IPv6 client.
@@ -4,7 +4,7 @@ Generated by `make aws-model-coverage`. This file is not published: it carries t

 **Not implemented** is the candidate work. **Not applicable** is declared in `pages.json` and is not work: those operations describe something this platform will not have.

-Pinned model source: `github.com/aws/api-models-aws@9b1a49292744, 2026-09-25`.
+Pinned model source: `github.com/aws/api-models-aws@e10d6a1c5ac3, 2026-10-28`.

 ## acm

@@ -46,7 +46,7 @@ Pinned model source: `github.com/aws/api-models-aws@9b1a49292744, 2026-09-25`.

 ## ec2

-`2016-11-15` — 127 of 804 modelled operations implemented (15.8%), 0 stubbed, 39 not applicable, 0 registered outside the pinned model.
+`2016-11-15` — 127 of 811 modelled operations implemented (15.7%), 0 stubbed, 39 not applicable, 0 registered outside the pinned model.

 ### Not implemented

@@ -190,6 +190,7 @@ Pinned model source: `github.com/aws/api-models-aws@9b1a49292744, 2026-09-25`.
 - `DeleteApplicationStatusCheck`
 - `DeleteCapacityManagerDataExport`
 - `DeleteClientVpnEndpoint`
+- `DeleteClientVpnEndpointAuthorizationPolicy`
 - `DeleteClientVpnRoute`
 - `DeleteCoipCidr`
 - `DeleteCoipPool`
@@ -484,7 +485,9 @@ Pinned model source: `github.com/aws/api-models-aws@9b1a49292744, 2026-09-25`.
 - `GetCapacityManagerMetricData`
 - `GetCapacityManagerMetricDimensions`
 - `GetCapacityManagerMonitoredTagKeys`
+- `GetCapacityManagerSupportabilityScores`
 - `GetCapacityReservationUsage`
+- `GetClientVpnEndpointAuthorizationPolicy`
 - `GetCoipPoolUsage`
 - `GetConsoleScreenshot`
 - `GetDeclarativePoliciesReportSummary`
@@ -496,6 +499,7 @@ Pinned model source: `github.com/aws/api-models-aws@9b1a49292744, 2026-09-25`.
 - `GetHostReservationPurchasePreview`
 - `GetImageAncestry`
 - `GetImageBlockPublicAccessState`
+- `GetImageFindings`
 - `GetInstanceMetadataDefaults`
 - `GetInstanceTpmEkPub`
 - `GetInstanceTypesFromInstanceRequirements`
@@ -532,6 +536,7 @@ Pinned model source: `github.com/aws/api-models-aws@9b1a49292744, 2026-09-25`.
 - `GetSnapshotBlockPublicAccessState`
 - `GetSpotPlacementScores`
 - `GetSubnetCidrReservations`
+- `GetTransitGatewayAttachmentNatConfiguration`
 - `GetTransitGatewayAttachmentPropagations`
 - `GetTransitGatewayMeteringPolicyEntries`
 - `GetTransitGatewayMulticastDomainAssociations`
@@ -563,6 +568,7 @@ Pinned model source: `github.com/aws/api-models-aws@9b1a49292744, 2026-09-25`.
 - `ModifyCapacityReservation`
 - `ModifyCapacityReservationFleet`
 - `ModifyClientVpnEndpoint`
+- `ModifyClientVpnEndpointAuthorizationPolicy`
 - `ModifyDefaultCreditSpecification`
 - `ModifyEbsDefaultKmsKeyId`
 - `ModifyFleet`
@@ -601,6 +607,7 @@ Pinned model source: `github.com/aws/api-models-aws@9b1a49292744, 2026-09-25`.
 - `ModifyTrafficMirrorFilterRule`
 - `ModifyTrafficMirrorSession`
 - `ModifyTransitGateway`
+- `ModifyTransitGatewayAttachmentNatConfiguration`
 - `ModifyTransitGatewayMeteringPolicy`
 - `ModifyTransitGatewayPolicyTableEntry`
 - `ModifyTransitGatewayPrefixListReference`
@@ -762,7 +769,7 @@ spot-price: On owned hardware there is no spot-to-on-demand price differential,

 ## ecr

-`2015-09-21` — 23 of 58 modelled operations implemented (39.7%), 10 stubbed, 0 not applicable, 3 registered outside the pinned model.
+`2015-09-21` — 23 of 62 modelled operations implemented (37.1%), 10 stubbed, 0 not applicable, 3 registered outside the pinned model.

 ### Not implemented

@@ -776,14 +783,18 @@ spot-price: On owned hardware there is no spot-to-on-demand price differential,
 - `DescribeImageReplicationStatus`
 - `DescribeImageSigningStatus`
 - `DescribePullThroughCacheRules`
+- `DescribeReplicationBackfillResults`
 - `DescribeRepositoryCreationTemplates`
 - `GetAccountSetting`
 - `GetSigningConfiguration`
 - `ListImageReferrers`
 - `ListPullTimeUpdateExclusions`
+- `ListReplicationBackfills`
 - `PutAccountSetting`
 - `PutSigningConfiguration`
 - `RegisterPullTimeUpdateExclusion`
+- `StartReplicationBackfill`
+- `StopReplicationBackfill`
 - `UpdateImageStorageClass`
 - `UpdatePullThroughCacheRule`
 - `UpdateRepositoryCreationTemplate`
@@ -819,19 +830,22 @@ spot-price: On owned hardware there is no spot-to-on-demand price differential,

 ## ecs

-`2014-11-13` — 31 of 77 modelled operations implemented (40.3%), 5 stubbed, 0 not applicable, 3 registered outside the pinned model.
+`2014-11-13` — 31 of 88 modelled operations implemented (35.2%), 5 stubbed, 0 not applicable, 3 registered outside the pinned model.

 ### Not implemented

+- `AssociateGatewayWithCluster`
 - `ContinueServiceDeployment`
 - `CreateDaemon`
 - `CreateExpressGatewayService`
+- `CreateGateway`
 - `CreateTaskSet`
 - `DeleteAccountSetting`
 - `DeleteAttributes`
 - `DeleteDaemon`
 - `DeleteDaemonTaskDefinition`
 - `DeleteExpressGatewayService`
+- `DeleteGateway`
 - `DeleteTaskDefinitions`
 - `DeleteTaskSet`
 - `DescribeDaemon`
@@ -839,9 +853,13 @@ spot-price: On owned hardware there is no spot-to-on-demand price differential,
 - `DescribeDaemonRevisions`
 - `DescribeDaemonTaskDefinition`
 - `DescribeExpressGatewayService`
+- `DescribeGateway`
+- `DescribeGatewayResources`
+- `DescribeGatewayRoute`
 - `DescribeServiceDeployments`
 - `DescribeServiceRevisions`
 - `DescribeTaskSets`
+- `DisassociateGatewayFromCluster`
 - `DiscoverPollEndpoint`
 - `ExecuteCommand`
 - `GetTaskProtection`
@@ -849,6 +867,9 @@ spot-price: On owned hardware there is no spot-to-on-demand price differential,
 - `ListDaemonDeployments`
 - `ListDaemonTaskDefinitions`
 - `ListDaemons`
+- `ListGatewayClusterAssociations`
+- `ListGatewayRoutes`
+- `ListGateways`
 - `ListServiceDeployments`
 - `PutAccountSettingDefault`
 - `PutAttributes`
@@ -861,6 +882,7 @@ spot-price: On owned hardware there is no spot-to-on-demand price differential,
 - `UpdateContainerAgent`
 - `UpdateDaemon`
 - `UpdateExpressGatewayService`
+- `UpdateGateway`
 - `UpdateServicePrimaryTaskSet`
 - `UpdateTaskProtection`
 - `UpdateTaskSet`
@@ -1085,7 +1107,7 @@ signing-certificates: X.509 signing certificates are an EC2-Classic SOAP credent

 ## rds

-`2014-10-31` — 27 of 164 modelled operations implemented (16.5%), 0 stubbed, 4 not applicable, 5 registered outside the pinned model.
+`2014-10-31` — 27 of 168 modelled operations implemented (16.1%), 0 stubbed, 4 not applicable, 5 registered outside the pinned model.

 ### Not implemented

@@ -1095,6 +1117,7 @@ signing-certificates: X.509 signing certificates are an EC2-Classic SOAP credent
 - `ApplyPendingMaintenanceAction`
 - `AuthorizeDBSecurityGroupIngress`
 - `BacktrackDBCluster`
+- `CancelDBCapacityReservation`
 - `CancelExportTask`
 - `CopyDBClusterParameterGroup`
 - `CopyDBClusterSnapshot`
@@ -1103,6 +1126,7 @@ signing-certificates: X.509 signing certificates are an EC2-Classic SOAP credent
 - `CopyOptionGroup`
 - `CreateBlueGreenDeployment`
 - `CreateCustomDBEngineVersion`
+- `CreateDBCapacityReservation`
 - `CreateDBClusterEndpoint`
 - `CreateDBClusterParameterGroup`
 - `CreateDBClusterSnapshot`
@@ -1133,6 +1157,7 @@ signing-certificates: X.509 signing certificates are an EC2-Classic SOAP credent
 - `DescribeAccountAttributes`
 - `DescribeBlueGreenDeployments`
 - `DescribeCertificates`
+- `DescribeDBCapacityReservations`
 - `DescribeDBClusterAutomatedBackups`
 - `DescribeDBClusterBacktracks`
 - `DescribeDBClusterEndpoints`
@@ -1174,6 +1199,7 @@ signing-certificates: X.509 signing certificates are an EC2-Classic SOAP credent
 - `ModifyCertificates`
 - `ModifyCurrentDBClusterCapacity`
 - `ModifyCustomDBEngineVersion`
+- `ModifyDBCapacityReservation`
 - `ModifyDBClusterEndpoint`
 - `ModifyDBClusterParameterGroup`
 - `ModifyDBClusterSnapshotAttribute`
@@ -1243,6 +1269,71 @@ option-groups: Option groups configure engine add-ons for engines this platform
 - `DescribeOptionGroups`
 - `ModifyOptionGroup`

+## route53
+
+`2013-04-01` — 13 of 71 modelled operations implemented (18.3%), 0 stubbed, 0 not applicable, 0 registered outside the pinned model.
+
+### Not implemented
+
+- `ActivateKeySigningKey`
+- `AssociateVPCWithHostedZone`
+- `ChangeCidrCollection`
+- `CreateCidrCollection`
+- `CreateHealthCheck`
+- `CreateKeySigningKey`
+- `CreateQueryLoggingConfig`
+- `CreateReusableDelegationSet`
+- `CreateTrafficPolicy`
+- `CreateTrafficPolicyInstance`
+- `CreateTrafficPolicyVersion`
+- `CreateVPCAssociationAuthorization`
+- `DeactivateKeySigningKey`
+- `DeleteCidrCollection`
+- `DeleteHealthCheck`
+- `DeleteKeySigningKey`
+- `DeleteQueryLoggingConfig`
+- `DeleteReusableDelegationSet`
+- `DeleteTrafficPolicy`
+- `DeleteTrafficPolicyInstance`
+- `DeleteVPCAssociationAuthorization`
+- `DisableHostedZoneDNSSEC`
+- `DisassociateVPCFromHostedZone`
+- `EnableHostedZoneDNSSEC`
+- `GetAccountLimit`
+- `GetCheckerIpRanges`
+- `GetDNSSEC`
+- `GetGeoLocation`
+- `GetHealthCheck`
+- `GetHealthCheckCount`
+- `GetHealthCheckLastFailureReason`
+- `GetHealthCheckStatus`
+- `GetHostedZoneLimit`
+- `GetQueryLoggingConfig`
+- `GetReusableDelegationSet`
+- `GetReusableDelegationSetLimit`
+- `GetTrafficPolicy`
+- `GetTrafficPolicyInstance`
+- `GetTrafficPolicyInstanceCount`
+- `ListCidrBlocks`
+- `ListCidrCollections`
+- `ListCidrLocations`
+- `ListGeoLocations`
+- `ListHealthChecks`
+- `ListHostedZonesByVPC`
+- `ListQueryLoggingConfigs`
+- `ListReusableDelegationSets`
+- `ListTrafficPolicies`
+- `ListTrafficPolicyInstances`
+- `ListTrafficPolicyInstancesByHostedZone`
+- `ListTrafficPolicyInstancesByPolicy`
+- `ListTrafficPolicyVersions`
+- `ListVPCAssociationAuthorizations`
+- `TestDNSAnswer`
+- `UpdateHealthCheck`
+- `UpdateHostedZoneFeatures`
+- `UpdateTrafficPolicyComment`
+- `UpdateTrafficPolicyInstance`
+
 ## s3

 `2006-03-01` — 26 of 112 modelled operations implemented (23.2%), 0 stubbed, 6 not applicable, 0 registered outside the pinned model.
@@ -247,6 +247,20 @@
        "option-groups": "Option groups configure engine add-ons for engines this platform does not offer, such as Oracle and SQL Server."
      }
    },
+    "route53": {
+      "slug": "route53",
+      "name": "Route 53",
+      "title": "Route 53 API Coverage",
+      "seoTitle": "Amazon Route 53 API Coverage on Spinifex — Spinifex Docs",
+      "description": "The Amazon Route 53 API operations Spinifex implements, covering the hosted zones it serves, their resource record sets, change status and tags on each zone.",
+      "tags": [
+        "aws",
+        "compatibility",
+        "coverage",
+        "route53",
+        "dns"
+      ]
+    },
    "s3": {
      "slug": "s3",
      "name": "S3",
@@ -0,0 +1,38 @@
+---
+title: "Route 53 API Coverage"
+seoTitle: "Amazon Route 53 API Coverage on Spinifex — Spinifex Docs"
+description: "The Amazon Route 53 API operations Spinifex implements, covering the hosted zones it serves, their resource record sets, change status and tags on each zone."
+category: "Coverage"
+sections:
+  - overview
+tags:
+  - aws
+  - compatibility
+  - coverage
+  - route53
+  - dns
+---
+
+# Route 53 API Coverage
+
+## Overview
+
+Spinifex implements **13 operations** in the Route 53 `2013-04-01` API model.
+
+### Operations
+
+| Operation |
+|---|
+| `ChangeResourceRecordSets` |
+| `ChangeTagsForResource` |
+| `CreateHostedZone` |
+| `DeleteHostedZone` |
+| `GetChange` |
+| `GetHostedZone` |
+| `GetHostedZoneCount` |
+| `ListHostedZones` |
+| `ListHostedZonesByName` |
+| `ListResourceRecordSets` |
+| `ListTagsForResource` |
+| `ListTagsForResources` |
+| `UpdateHostedZoneComment` |
@@ -163,11 +163,14 @@ Open between all hosts:
 | Protocol and port | Used by |
 |---|---|
 | UDP 6081 | Geneve tunnels |
+| UDP 500, 4500 and IP protocol 50 (ESP) | IPsec for the Geneve tunnels, while IPsec is on |
 | TCP 4222, 4248 | NATS |
 | TCP 6641, 6642 | OVN northbound and southbound |
 | TCP 8443 | Predastore S3 gate |
 | UDP 6660, 7660 | Object shards, metadata consensus |

+UDP 4500 carries the tunnel traffic only where IKE detects NAT between two nodes; otherwise it travels as ESP, on Oracle Linux as on Debian.
+
 Open between the three OVN database servers (servers 1 to 3) only:

 | Protocol and port | Used by |
@@ -207,7 +210,10 @@ Spinifex's own **host** firewall is separate, and already carries this policy

 Choose one method and apply it to **every** server in the cluster.

-**Option A — existing OS.** On a server already running Ubuntu 26.04 or Debian 13:
+**Option A — existing OS.** Use this on servers already running Ubuntu 26.04, Debian 13 or Oracle Linux 10, and prepare an Oracle Linux server with [Oracle Linux Hosts](/docs/install-oracle-linux) first.
+Every server in a cluster must run the same OS: Oracle Linux never shares a cluster with Debian or Ubuntu, and `spx admin join` does not check this for you.
+
+Run the installer on each server:

 ```bash
 curl -fsSL https://install.mulgadc.com | bash
@@ -0,0 +1,270 @@
+---
+title: "Oracle Linux Hosts"
+seoTitle: "Install Spinifex on Oracle Linux 10 — Spinifex Docs"
+description: "Prepare Oracle Linux 10 hosts for Spinifex: UEK R8, SELinux, a NetworkManager br-wan bridge, the NFV SIG repository, IPsec on libreswan and host updates."
+category: "Install"
+tags:
+  - install
+  - oracle linux
+  - single node
+  - multi node
+resources:
+  - title: "Oracle Linux 10 Documentation"
+    url: "https://docs.oracle.com/en/operating-systems/oracle-linux/10/"
+  - title: "CentOS NFV SIG"
+    url: "https://sigs.centos.org/nfv/"
+---
+
+# Oracle Linux Hosts
+
+> What differs when the Spinifex host runs Oracle Linux 10 instead of Debian or Ubuntu.
+
+## Table of Contents
+
+- [Overview](#overview)
+- [Prerequisites](#prerequisites)
+- [Instructions](#instructions)
+- [Troubleshooting](#troubleshooting)
+
+---
+
+## Overview
+
+Spinifex installs on Oracle Linux 10.2 or a later 10.x release, running the UEK R8 kernel, on x86_64.
+The installer, the release tarball and the install steps are the same as on Debian; this page covers what the host needs first and what the installer changes on it.
+Follow it alongside [Single-Node Install](/docs/install).
+
+**Supported on Oracle Linux:**
+
+- A single node, or a cluster in which every node runs Oracle Linux; servers beyond the third join as compute nodes, as on Debian.
+- Guests, VPCs, routed NAT, Elastic IPs, and Application and Network Load Balancers.
+- Whole-GPU passthrough, on x86_64 GPU hosts (see [GPU Passthrough](/docs/gpu-passthrough)).
+
+**Not supported on Oracle Linux yet:**
+
+- Clusters that mix Oracle Linux with Debian or Ubuntu nodes, on any release.
+- SELinux enforcing, aarch64, Oracle Linux 9 or earlier, the Red Hat Compatible Kernel, a ZFS root, and NVIDIA MIG.
+
+Open vSwitch and OVN come from the CentOS NFV SIG, because Oracle ships neither for Oracle Linux 10.
+The NFV SIG is a community repository, and Oracle does not support the packages it publishes.
+Spinifex installs `openvswitch3.5` and `ovn25.03`, the versions Debian 13 ships, and its repository file lets the SIG supply nothing else.
+QEMU comes from Oracle's KVM Utilities repository, because only that build has the `microvm` machine type Spinifex load balancers boot on.
+
+An Oracle Linux node must be a dedicated Spinifex host.
+Spinifex claims every `tap*` device and owns libreswan's `ipsec.service`, which it keeps stopped on a single node and runs for its own tunnels on a cluster node, so VPNs, hand-run QEMU guests and libreswan tunnels of your own cannot share the host.
+Crypto policies other than `DEFAULT` and `FIPS` are untested, and so is Secure Boot.
+
+## Prerequisites
+
+### Operating system and kernel
+
+Install Oracle Linux 10 with Oracle's ISO or a KVM template, then update it to the current minor release and reboot:
+
+```bash
+sudo dnf update -y
+sudo reboot
+```
+
+After the reboot, check the release, the kernel and the CPU:
+
+```bash
+cat /etc/oracle-release        # Oracle Linux Server release 10.2 or later
+uname -r                       # 6.12.0-...el10uek.x86_64
+grep -c -w avx2 /proc/cpuinfo  # non-zero
+```
+
+The kernel must be UEK R8 (`6.12.0-*uek`), and it must be the running kernel: an installed but not booted `kernel-uek` is refused.
+The CPU must be x86-64-v3 or newer (Intel Haswell, AMD Excavator or later), since the NFV SIG and KVM Utilities builds target it.
+The installer names any missing CPU flag.
+
+The installer adds `kernel-uek-modules` for the running kernel, which carries `openvswitch.ko` and pulls in `linux-firmware`, about 800 MB.
+Oracle publishes modules for each kernel separately, so a host running an old kernel may find its modules gone from the repository; the installer then stops and asks for a kernel update and a reboot.
+
+### SELinux
+
+Spinifex ships no SELinux policy yet, so SELinux must be permissive or disabled, both now and at the next boot.
+Oracle's images boot enforcing.
+
+```bash
+sudo setenforce 0
+sudo sed -i 's/^SELINUX=.*/SELINUX=permissive/' /etc/selinux/config
+getenforce                     # Permissive
+```
+
+The installer and `spx admin preflight` refuse a host that is enforcing in either place.
+
+### WAN bridge with NetworkManager
+
+As on every OS, the host's uplink must already sit on a Linux bridge named `br-wan`, with the host address and default route on the bridge (see [Single-Node Install](/docs/install#prerequisites)).
+On Oracle Linux, NetworkManager owns `br-wan`; Spinifex leaves the bridge and its NIC to it and marks the links it creates itself as unmanaged.
+
+Create the bridge from the console, not over SSH on the NIC being moved, since the host loses its address while the profile changes.
+Replace `eno1` with the uplink NIC and `"System eno1"` with its current profile (`nmcli connection show`):
+
+```bash
+sudo nmcli connection add type bridge ifname br-wan con-name br-wan \
+  bridge.stp no ipv4.method auto ipv6.method disabled
+sudo nmcli connection add type bridge-slave ifname eno1 master br-wan con-name br-wan-port
+sudo nmcli connection down "System eno1"
+sudo nmcli connection up br-wan
+```
+
+For a static address, give the bridge `ipv4.method manual` with `ipv4.addresses`, `ipv4.gateway` and `ipv4.dns` instead of `ipv4.method auto`.
+
+`bridge.stp no` matters: a NetworkManager bridge runs STP by default, which holds the port in its forwarding delay for about 15 seconds after every link up, long enough for DHCP on the bridge to time out.
+Then verify, as on Debian:
+
+```bash
+ip -br link show br-wan        # UP
+ip route                       # default via ... dev br-wan
+```
+
+### Network access during install
+
+Besides `install.mulgadc.com` and the AWS CLI download, the install reaches `yum.oracle.com` for the Oracle Linux repositories, `mirror.stream.centos.org` for the NFV SIG packages, and `www.centos.org` for the NFV SIG signing key.
+An air-gapped install on Oracle Linux is not documented yet.
+
+## Instructions
+
+## Step 1. Install Spinifex
+
+Run the installer as a user with `sudo`, exactly as on Debian:
+
+```bash
+curl -fsSL https://install.mulgadc.com | bash
+```
+
+For a single node, continue with [Single-Node Install](/docs/install) from Step 2: `setup-ovn.sh --management`, `spx admin init` and `systemctl start spinifex.target`.
+IPsec stays off on a single node, because there is no peer to encrypt traffic to.
+
+For a cluster, prepare and install every node this way, then continue with [Multi-Node Install](/docs/install-multi-node) from Step 2.
+
+### Clusters
+
+Every node in an Oracle Linux cluster runs Oracle Linux.
+`install-node.sh` refuses a host list that mixes Oracle Linux with Debian or Ubuntu, but `spx admin join` does not check, so keep to one OS yourself when joining by hand.
+
+IPsec encrypts the Geneve tunnels between nodes and is on by default, as on Debian.
+On Oracle Linux it runs on libreswan, the IKE daemon Oracle ships, rather than strongSwan.
+Once the node's IPsec certificate exists, the daemon unmasks and starts `ipsec.service` and `openvswitch-ipsec.service`.
+`ovs-monitor-ipsec` then loads the node certificate, its key and the cluster CA into libreswan's NSS database at `/var/lib/ipsec/nss`, under nicknames that start with `ovs_cert`.
+
+The tunnels are plain ESP (IP protocol 50), as on Debian, negotiated over UDP 500.
+Open UDP 500, UDP 4500 and ESP between nodes: libreswan moves to UDP 4500 if it detects NAT between two nodes.
+The Spinifex host firewall opens all three to cluster peers; a firewall between the nodes must allow them too.
+While IKE to a peer fails, the node drops its Geneve traffic to that peer rather than send it in clear.
+The nft table `inet spinifex_ipsec_guard` drops outbound UDP 6081 from the encap address whose route carries no SA, which also covers the moments pluto flushes the xfrm policy table as it starts and stops.
+It is not persisted: `ipsec.service` puts it back at boot before ovn-controller builds the tunnels, and the daemon re-syncs it every 60 seconds.
+`spx admin init --ipsec=false` leaves the tunnels unencrypted, as on any OS.
+
+## Step 2. Know What the Installer Changed
+
+On Oracle Linux the installer makes these changes, which Debian hosts do not need:
+
+- **Repositories.** It enables `ol10_UEKR8` and `ol10_kvm_utils`, and adds `qemu-kvm*` and `qemu-img*` to the `excludepkgs` of `ol10_appstream`, keeping any excludes already set, so a later AppStream build can never replace the KVM Utilities QEMU.
+- **NFV SIG.** It writes `/etc/yum.repos.d/spinifex-nfv-sig.repo` with an `includepkgs=` list of the OVS and OVN packages, and imports the SIG key to `/etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-SIG-NFV` only when its fingerprint is `35154228174901BEFA8E69A621465E289D2A76A7`.
+- **OVS and OVN user.** Both run as root, as on Debian, through `OVS_USER_ID` and `OVN_USER_ID` in `/etc/sysconfig/openvswitch` and `/etc/sysconfig/ovn`.
+- **DNS.** Where systemd-resolved is installed but not running, it saves `/etc/resolv.conf` to `/var/lib/spinifex-reset/resolv.conf.orig`, starts resolved behind NetworkManager and links `/etc/resolv.conf` to the resolved stub. It then checks that names still resolve and stops, printing the commands that put DNS back, if they do not.
+- **chrony.** It appends `confdir /etc/chrony.d` to `/etc/chrony.conf` once, and writes its step policy to `/etc/chrony.d/spinifex.conf`.
+- **`spx` under `sudo`.** Oracle Linux's `sudo` path leaves out `/usr/local/bin`, so the installer links `/usr/bin/spx` to `/usr/local/bin/spx`.
+
+The kernel modules list (`/etc/modules-load.d/spinifex.conf`) and the host firewall below apply on every OS.
+
+## Step 3. Replace firewalld Rules You Depend On
+
+The installer stops and masks firewalld, because with it running a packet must pass both firewalld's table and the Spinifex one, which silently breaks routed NAT, Elastic IPs and instance metadata.
+When firewalld was active, the installer turns on the Spinifex nftables policy in its place; pass `--firewall=on` or `--firewall=off` to choose yourself.
+
+firewalld's rules are not saved or migrated.
+The Spinifex policy filters inbound traffic only and opens 53, 443, 3000, 8443 and 9999 to any source.
+It accepts SSH from every IPv4 source until `$trusted_ssh_peers` in `/etc/spinifex/firewall/custom.nft` narrows it.
+Restate any other restriction you had, such as SSH limited to a management subnet, in that file; [Host Firewall](/docs/host-firewall) shows how.
+
+From the moment firewalld stops until the daemon first loads the Spinifex policy, the host accepts every connection.
+Uninstalling Spinifex leaves firewalld unmasked but disabled; it does not turn it back on.
+
+## Step 4. Keep the Host Updated
+
+Do not enable `dnf-automatic`.
+An Open vSwitch package upgrade restarts OVS, which interrupts guest networking, so update `openvswitch3.5`, `ovn25.03*` and `qemu-kvm*` in a maintenance window.
+Stay on the current Oracle Linux 10 minor release, which is what Spinifex is tested on.
+
+Before rebooting into a new kernel, check that its modules are installed, or OVS cannot start after the reboot:
+
+```bash
+rpm -q kernel-uek-core kernel-uek-modules
+```
+
+Each `kernel-uek-core` version needs a `kernel-uek-modules` of the same version; install a missing one with `sudo dnf install kernel-uek-modules-<version>`.
+
+Track Oracle Linux security advisories (ELSA) for the host, the kernel and QEMU, for example with `dnf updateinfo list --security`.
+Those advisories do not cover the NFV SIG packages, so follow Open vSwitch and OVN security announcements for those.
+
+## Step 5. Uninstall
+
+`/usr/local/share/spinifex/uninstall-spx.sh` removes Spinifex as on Debian.
+On a cluster node it also stops `openvswitch-ipsec.service` and deletes the `ovs_cert*` entries from libreswan's NSS database, so the node can join a different cluster; other NSS identities and your connection files in `/etc/ipsec.d` stay.
+It lists what it leaves in place on Oracle Linux: the repository changes and the SIG key, `OVS_USER_ID` and `OVN_USER_ID`, the OVS and OVN logrotate edits, the chrony include, `/etc/modules-load.d/spinifex.conf`, the SELinux mode, a masked `ipsec.service`, firewalld disabled, and any GPU passthrough configuration.
+`--purge-deps` is refused on Oracle Linux, because dnf cannot tell the packages Spinifex installed from the host's own; remove them with `dnf remove` yourself.
+
+## Troubleshooting
+
+### Installer stops: SELinux is enforcing
+
+Set SELinux permissive now and in `/etc/selinux/config`, as in [SELinux](#selinux), then run the installer again.
+Fixing only one of the two is refused too.
+
+### Installer stops: the running kernel is not UEK R8
+
+Install `kernel-uek`, reboot into it, check `uname -r`, then run the installer again.
+
+### Installer stops: kernel-uek-modules is no longer in the repositories
+
+The running kernel is older than anything the repository still carries modules for.
+
+```bash
+sudo dnf update -y kernel-uek
+sudo reboot
+```
+
+### Installer stops: a repository is unreachable
+
+The message names the repository.
+Check the host can reach `yum.oracle.com` and `mirror.stream.centos.org` over HTTPS, including through any proxy dnf uses, then run the installer again.
+
+### DHCP on br-wan times out
+
+Check STP is off on the bridge:
+
+```bash
+nmcli -g bridge.stp connection show br-wan
+```
+
+If it prints `yes`, run `sudo nmcli connection modify br-wan bridge.stp no` and `sudo nmcli connection up br-wan`.
+
+### `sudo spx` finds no command
+
+Check that `/usr/bin/spx` links to `/usr/local/bin/spx`; run the installer again if it does not, or call `sudo /usr/local/bin/spx` meanwhile.
+
+### Cluster nodes form no IPsec tunnels
+
+Check that both IPsec services run, and that libreswan holds the node's certificate and the cluster CA:
+
+```bash
+systemctl is-active ipsec.service openvswitch-ipsec.service
+sudo certutil -L -d sql:/var/lib/ipsec/nss     # ovs_certkey_<node>, ovs_cert_cacert
+sudo ipsec trafficstatus                       # one line per peer tunnel
+```
+
+If the services are inactive, check the daemon log with `journalctl -u spinifex-daemon` for the IPsec reconcile.
+If the certificates are present but no tunnel is listed, check that UDP 500 and ESP pass between the nodes in both directions, and look for `SEC_ERROR_UNKNOWN_ISSUER` in `journalctl -u ipsec`, which means a peer's certificate is from another CA.
+Until the tunnel forms, guest traffic between those two nodes is dropped, which is the guard in [Clusters](#clusters) working.
+
+### CA certificate not trusted
+
+`sudo spx admin init` installs the CA automatically. To install it again by hand on Oracle Linux:
+
+```bash
+sudo cp /etc/spinifex/ca.pem /etc/pki/ca-trust/source/anchors/spinifex-ca.crt
+sudo update-ca-trust
+```
@@ -1,7 +1,7 @@
 ---
 title: "Source Install"
 seoTitle: "Build and Install Spinifex from Source — Spinifex Docs"
-description: "Build Spinifex from source on Ubuntu or Debian for development, custom builds, or contributing changes, then install and run the resulting binaries locally."
+description: "Build Spinifex from source on Debian, Ubuntu or Oracle Linux for development, custom builds or contributing changes, then install and run the binaries."
 category: "Install"
 tags:
  - install
@@ -34,6 +34,7 @@ This guide builds Spinifex from source. For production deployments, the [binary

 - Ubuntu 26.04
 - Debian 13
+- Oracle Linux 10.2 or a later 10.x, on UEK R8, x86_64 only. Prepare the host with [Oracle Linux Hosts](/docs/install-oracle-linux) first; `make quickinstall` also enables the CodeReady Builder repository for `nbdkit-devel`.

 ## Instructions

@@ -1,7 +1,7 @@
 ---
 title: "Single-Node Install"
 seoTitle: "Install Spinifex on a Single Server — Spinifex Docs"
-description: "Install Spinifex on one Ubuntu or Debian server with the binary installer and get an AWS-compatible EC2, S3, and VPC stack running on your own hardware."
+description: "Install Spinifex on one Debian, Ubuntu or Oracle Linux server with the binary installer and run an AWS-compatible EC2, S3 and VPC stack on your hardware."
 category: "Install"
 tags:
  - install
@@ -39,6 +39,7 @@ This guide installs Spinifex on a single server using the binary installer. For

 - Ubuntu 26.04
 - Debian 13
+- Oracle Linux 10.2 or a later 10.x, on UEK R8, x86_64 only, as a single node or, with [Multi-Node Install](/docs/install-multi-node), a cluster. Open vSwitch and OVN come from the CentOS NFV SIG, a community repository that Oracle does not support. Prepare the host with [Oracle Linux Hosts](/docs/install-oracle-linux) first.

 **What Gets Installed:**

@@ -139,19 +140,33 @@ Continue to [Setting Up Your Cluster](/docs/setting-up-your-cluster) to import a

 ### spx Command Not Found

+The installer puts `spx` in `/usr/local/bin`.
+Where `sudo`'s `secure_path` leaves that directory out, as on Oracle Linux, the installer also links `/usr/bin/spx` to it, so `sudo spx` works; changing `PATH` in your shell does not reach commands run with `sudo`.
+Check both:
+
 ```bash
-export PATH=$PATH:/usr/local/bin
+ls -l /usr/local/bin/spx /usr/bin/spx
+sudo sh -c 'command -v spx'
 ```

+If `sudo` still finds no `spx`, run the installer again, or call `sudo /usr/local/bin/spx` meanwhile.
+
 ### CA Certificate Not Trusted

-`sudo spx admin init` installs the CA automatically. If you need to re-install it manually:
+`sudo spx admin init` installs the CA automatically. If you need to re-install it manually on Debian or Ubuntu:

 ```bash
 sudo cp /etc/spinifex/ca.pem /usr/local/share/ca-certificates/spinifex-ca.crt
 sudo update-ca-certificates
 ```

+On Oracle Linux:
+
+```bash
+sudo cp /etc/spinifex/ca.pem /etc/pki/ca-trust/source/anchors/spinifex-ca.crt
+sudo update-ca-trust
+```
+
 ### OVN Services Not Starting

 ```bash
@@ -269,6 +269,35 @@ ip link set <dev> mtu 9000        # both ends
 ping -M do -s 8972 -c 3 <peer>    # must succeed
 ```

+### The internet path
+
+A guest MTU of 1408 or 1442 sits below the 1500 bytes the internet carries, so a server can send the guest segments it cannot receive.
+The VPC gateway port carries `gateway_mtu = min(guest MTU, 1500)` and answers an oversized packet with ICMP fragmentation-needed (IPv4) or Packet Too Big (IPv6).
+It follows `ipsec_enabled` and `network.underlay_mtu` and converges on the next reconcile pass, with no VPC recreate.
+
+Path MTU discovery depends on that ICMP reaching the server, and an upstream NAT router or a CDN often drops it.
+vpcd therefore also clamps the TCP MSS on the host hop to the gateway MTU minus 40 (IPv4) or minus 60 (IPv6), and re-asserts the rule every 30 seconds so a firewall reload cannot remove it for long.
+
+| Uplink | Clamp | Table |
+| --- | --- | --- |
+| veth (distributed NAT) | yes | `bridge spinifex_mss`, matching the `veth-wan-br` port |
+| routed NAT | yes | `inet spinifex_mss`, matching `spx-nat-host` |
+| physical | no host hop exists | path MTU discovery only |
+
+On a physical uplink the only cover is path MTU discovery, which fails when the upstream drops the ICMP or the peer ignores it.
+
+Measured on one node (env12), IPsec on, Ubuntu 26.04 guest, 16 MB HTTPS download inside a Docker container at the default bridge MTU of 1500:
+
+| Uplink | Without the clamp | With the clamp |
+| --- | --- | --- |
+| veth, instance with a public IP | stalled 3/3 | completed 3/3 |
+| routed NAT | stalled 3/3, no ICMP seen on the host | completed 3/3 |
+
+An instance behind a NAT gateway on the veth uplink also completed 3/3 with the clamp; its unclamped internet run was not measured.
+The guest's own `curl` completed in every case.
+Without the clamp, `gateway_mtu` alone fixed the download only against a server on the same L2 with no NAT in between (completed 3/3), and not against the internet through the lab's upstream NAT router.
+Not measured: a physical uplink, a multi-node cluster, IPv6, and a Debian guest.
+
 ## 5. Measurement notes

 Throughput on this stack is **noisy** — repeated identical 20s runs varied between
@@ -127,8 +127,8 @@ Operators own remediation for layers Spinifex does not ship:
 | Layer | Responsibility |
 |-------|----------------|
 | Spinifex releases | Apply released patches within the correct-SLA above. |
-| Host OS + kernel | Track distribution advisories (Debian DSA, Ubuntu USN) at matching cadence. |
-| OVN / OVS | Patched via the operator's package channel alongside the OS. |
+| Host OS + kernel | Track distribution advisories (Debian DSA, Ubuntu USN, Oracle Linux ELSA) at matching cadence. |
+| OVN / OVS | Patched via the operator's package channel alongside the OS. On Oracle Linux they come from the CentOS NFV SIG, which ELSA does not cover, so track upstream Open vSwitch and OVN security announcements. |
 | QEMU / KVM / libvirt | As above — hypervisor CVEs affect tenant isolation. |
 | Host AV / EDR agent | See [Malware Protection §2](/docs/malware-protection#2-update-requirements-sil1-3144). |

@@ -75,7 +75,7 @@ All Spinifex nodes must include the following in their malware protection scope.
 | `/usr/local/bin/spx` | Spinifex CLI and service entry point. All `spinifex-*.service` units invoke `spx service <name> start`, so every service binary is reached through this one path. |
 | `/usr/bin/ovn-*`, `/usr/bin/ovs-*` | OVN / OVS control-plane and datapath binaries (installed via `apt` from the `ovn-central`, `ovn-host`, `openvswitch-switch` packages). |
 | `/usr/lib/<arch>-linux-gnu/nbdkit/plugins/nbdkit-viperblock-plugin.so` | Viperblock NBD plugin loaded by `nbdkit` to expose EBS volumes as block devices. The exact path comes from `nbdkit --dump-config` (`plugindir=`); on Debian 13 amd64 it is `/usr/lib/x86_64-linux-gnu/nbdkit/plugins/`. |
-| `/usr/local/share/ca-certificates/spinifex-ca.crt` | Cluster CA cert installed into the system trust store |
+| `/usr/local/share/ca-certificates/spinifex-ca.crt` | Cluster CA cert installed into the system trust store. On Oracle Linux it is `/etc/pki/ca-trust/source/anchors/spinifex-ca.crt`. |

 ### 1.2 Configuration and Secrets

@@ -108,8 +108,8 @@ behavior in other words.
 | 6643 | OVN Northbound DB (RAFT) | OVSDB/TCP | Cluster | NB database RAFT replication between the 3 quorum nodes | Cluster network only; TLS planned |
 | 6644 | OVN Southbound DB (RAFT) | OVSDB/TCP | Cluster | SB database RAFT replication between the 3 quorum nodes | Cluster network only; TLS planned |
 | 6081 | OVN (Geneve) | UDP | Encap | Tenant traffic overlay between chassis. A kernel UDP-tunnel socket, so packets are delivered locally and traverse the host's netfilter input hook before OVS sees them — a host firewall must accept them explicitly. The socket is opened by the kernel tunnel driver and takes no bind address, so it binds the wildcard by design and reach must be restricted by firewall. | None — see 500/4500 |
-| 500, 4500 | strongSwan `charon` | IKEv2 / UDP | Encap | IKE and NAT-T for the IPsec protecting Geneve, managed entirely by `ovs-monitor-ipsec`. Binds the wildcard by design (the upstream strongSwan default, accepted rather than overridden), so reach must be restricted by firewall. | Certificate-based, against the cluster CA |
-| — | ESP | IP proto 50 | Encap | The IPsec payload itself, once IKE has negotiated an SA | Cluster CA |
+| 500, 4500 | strongSwan `charon`, or libreswan `pluto` on Oracle Linux | IKEv2 / UDP | Encap | IKE and NAT-T for the IPsec protecting Geneve, managed entirely by `ovs-monitor-ipsec`; 4500 carries the ESP payload only where IKE detects NAT between two nodes. `charon` binds the wildcard by design (the upstream strongSwan default, accepted rather than overridden); `pluto` binds each host address separately, the WAN address included, which is no narrower. Either way reach must be restricted by firewall. | Certificate-based, against the cluster CA |
+| — | ESP | IP proto 50 | Encap | The IPsec payload itself, once IKE has negotiated an SA, on Oracle Linux as on Debian. | Cluster CA |
 | 8222 | spinifex-nats (monitoring) | HTTP | Localhost | `varz`/`subsz` metrics consumed by the daemon | Loopback only |
 | 323 | chronyd | NTP | Localhost | Time sync client control socket | Loopback only |
 | 169.254.169.254:80 | spinifex-vpcd (IMDS) | HTTP | Guest | Instance metadata service. One socket per instance, bound to that instance's `ime-*` interface. Terminates on the host, so it traverses the netfilter input hook. | Instance identity by interface; IMDSv2 tokens |
@@ -145,6 +145,16 @@ Spinifex nodes initiate a small, fixed set of outbound connections.
 | `https://d2yp8ipz5jfqcw.cloudfront.net` | Alpine image for managed HAProxy load-balancer | HTTPS | TLS + checksum verification |
 | `https://install.mulgadc.com/install` | One-shot install telemetry POST on `spx admin init` / `join`. | HTTPS | TLS |

+**During install and package updates on Oracle Linux hosts:**
+
+| Destination | Purpose | Protocol | Verification |
+|-------------|---------|----------|--------------|
+| `https://yum.oracle.com` | Oracle Linux BaseOS, AppStream, UEK R8 and KVM Utilities repositories | HTTPS | TLS + Oracle RPM signatures |
+| `https://mirror.stream.centos.org/SIGs/10-stream/nfv/` | Open vSwitch and OVN from the CentOS NFV SIG, limited by `includepkgs=` | HTTPS | TLS + NFV SIG RPM signatures |
+| `https://www.centos.org/keys/RPM-GPG-KEY-CentOS-SIG-NFV` | NFV SIG signing key, fetched once at install | HTTPS | TLS + fingerprint pinned in `setup.sh` |
+
+Debian and Ubuntu hosts reach their distribution's own mirrors for the same purpose.
+
 **To peer nodes (cluster-internal):** NATS federation (4248), Predastore S3 (8443), OVN NB/SB (6641/6642), northstar DNS forwarding (5300), and the Geneve/IPsec overlay (6081, 500, 4500, ESP) — see [§3](#3-cross-node-internal-connections) for encryption and verification of each. The daemon also polls local NATS monitoring at `127.0.0.1:8222/varz` (loopback HTTP). It opens no connection to Predastore for status: the storage topology it reports comes from reading `predastore.toml`, and no Predastore node serves a status endpoint.

 **Update checks and metadata.** Spinifex does not check for updates and does not consume a cloud metadata service (`169.254.169.254` is served *by* the cluster to guest VMs). Node software updates come from the operator's OS package channel. The install-telemetry endpoint above is the only vendor-operated destination contacted by a node; closed-egress deployments should disable it and record the opt-out in the security plan.
@@ -164,7 +174,7 @@ Control-plane and data-plane traffic between Spinifex nodes, for completeness an
 | OVN NB/SB (client) | 6641 / 6642 | Cluster network only (TLS planned) | Network control plane; vpcd and ovn-controller dial the quorum |
 | OVN NB/SB (RAFT) | 6643 / 6644 | Cluster network only (TLS planned) | NB/SB database replication across the 3 quorum nodes |
 | OVN tunnels (Geneve) | UDP 6081 | Encapsulated by IPsec when `network.ipsec_enabled` is true, which is the default on multi-node clusters | Tenant traffic overlay between chassis, on the `vpc` plane |
-| OVN IPsec (IKE / NAT-T / ESP) | UDP 500, UDP 4500, IP proto 50 | Certificate-based against the cluster CA, negotiated by `ovs-monitor-ipsec` | Protects the Geneve tunnels above. OVN-native only — no layer manages strongSwan directly. |
+| OVN IPsec (IKE / NAT-T / ESP) | UDP 500, UDP 4500, IP proto 50 | Certificate-based against the cluster CA, negotiated by `ovs-monitor-ipsec` | Protects the Geneve tunnels above. OVN-native only: no layer configures strongSwan or libreswan directly. On Oracle Linux the daemon also blocks outbound Geneve that no SA covers, so a peer failing IKE gets no plaintext. |
 | Instance DNS forwarding | UDP/TCP 5300 | None | Each node's per-instance DNS shim forwards guest queries to peer nodes' northstar `:5300` |

 Nodes **must** sit on a network segment that is not routed to tenant/guest VMs or to the internet. The Predastore blob and meta transports and the OVN DBs are cluster-internal and must not be reachable from anywhere else.
@@ -173,7 +183,10 @@ Nodes **must** sit on a network segment that is not routed to tenant/guest VMs o

 Default external surface is five listeners — **9999** (AWS API), **3000** (UI), **22** (SSH), **8443** (S3) and **53** (DNS) — plus **4432** transiently during bootstrap. Every other listener is cluster- or encap-scoped and the operator must enforce this with a host firewall or an upstream network ACL.

-The nodes ship no firewall policy today. Until one is installed, this is the operator's responsibility and the reference below is the starting point.
+Every node ships a host firewall policy, `table inet spinifex_filter` in `/etc/spinifex/firewall/spinifex.nft`, written by `setup.sh`; `spinifex-daemon` loads it once cluster peers resolve, filling the peer and encap sets itself.
+It is armed by default on ISO installs; a `setup.sh` install arms it with `--firewall=on`, or when firewalld was active, and otherwise installs it unarmed, so on those nodes the scope split is still the operator's to enforce.
+`setup.sh` masks firewalld wherever it is installed, and `spx admin preflight` fails while firewalld or ufw is active, because a second ruleset silently drops routed NAT, elastic IP and metadata traffic.
+The shipped policy is the authority; the reference below explains the rules it needs and is looser than the shipped one in places, such as accepting all `ime-*` traffic where the shipped policy accepts only the metadata and DNS endpoints.

 > **Read the notes under the ruleset before applying it.** A default-deny input policy that
 > omits any of the loopback, conntrack, `ime-*` or Geneve rules will break instance
@@ -223,6 +236,7 @@ Notes that make the difference between this working and locking you out:

 - **Use a dedicated table and never flush the others.** `vpcd` writes MASQUERADE and per-EIP FORWARD rules into the `ip filter` and `ip nat` tables, and only reinstalls them when the service starts. `iptables -F`, `nft flush ruleset`, `ufw enable` and `firewalld` all destroy them silently, and instance networking stays broken until the next restart.
 - **Do not add a `forward` hook.** nftables evaluates every table registered on a hook and any `drop` is final, so a default-deny forward chain here would override `vpcd`'s accepts in the other table and break every routed-NAT instance and every EIP. Filtering forwarded guest traffic is the security group's job, not the host firewall's.
+- **`vpcd` owns a second table named `spinifex_mss`, separate from `inet spinifex_filter`.** It is a `bridge` table on the veth uplink and an `inet` table on the routed uplink, it has a `forward` hook with policy `accept`, and it only rewrites the MSS option of TCP SYN packets on the host hop to the gateway. It never drops or accepts traffic and does not replace the forward-hook advice above. `vpcd` re-installs it every 30 seconds, so a flush is repaired within that interval, and there is none on a physical uplink.
 - **`output` is deliberately untouched.** The metadata service's reply path egresses through per-instance policy routing; filtering `output` breaks it in ways that are hard to attribute.
 - **On a single-NIC node the cluster and encap sets are the node's public addresses**, because the planes collapse. The rules are still correct, but they are no longer a boundary — an upstream ACL is the only real control there.

@@ -234,7 +248,7 @@ Every listener and outbound destination is controlled by one of these files. Cha

 | File | Keys | Controls |
 |------|------|----------|
-| `/etc/spinifex/spinifex.toml` | `nodes.<node>.{awsgw,nats,predastore,daemon}.host`, `nodes.<node>.vpcd.ovn_{nb,sb}_addr`, `nodes.<node>.daemon.dev_networking`, `network.ipsec_enabled` | Per-service bind addresses/ports; dev-mode QEMU port forwarding. `network.ipsec_enabled` (default `true`) decides whether OVN-native IPsec protects the Geneve tunnels, and therefore whether `charon` listens on 500/4500 — single-node clusters never enable it. |
+| `/etc/spinifex/spinifex.toml` | `nodes.<node>.{awsgw,nats,predastore,daemon}.host`, `nodes.<node>.vpcd.ovn_{nb,sb}_addr`, `nodes.<node>.daemon.dev_networking`, `network.ipsec_enabled` | Per-service bind addresses/ports; dev-mode QEMU port forwarding. `network.ipsec_enabled` (default `true`) decides whether OVN-native IPsec protects the Geneve tunnels, and therefore whether the IKE daemon (`charon`, or `pluto` on Oracle Linux) listens on 500/4500 — single-node clusters never enable it. |
 | `/etc/spinifex/nats.conf` | `listen`, `cluster.listen`, `cluster.routes`, `http`, `tls`, `cluster.authorization` | NATS client/cluster/monitoring listeners, peer routes, TLS, cluster token. |
 | `/etc/spinifex/predastore/predastore.toml` | `[[host]].bind_addr`, `[[host]].addr`, `[[host]].tls_cert`, `[[host]].tls_key`, `[[host.node]].role`, `[[host.node]].port` | Predastore host and node layout: `bind_addr` is the address the host's sockets bind, `addr` is the address peer hosts dial it on, and both carry no port — the nodes pinned to the host supply their own. The service's `--host`/`--port` override the bind address and the gate's S3 port. |
 | `/etc/spinifex/northstar/northstar.toml` | `listen`, `forwarders` | DNS listen addresses (`:53` on the advertise address, `:5300` wildcard) and upstream resolvers. |
@@ -1,521 +0,0 @@
-# Service interface manifest.
-#
-# Source of truth for "which services produce/consume which NATS subjects"
-# and "which e2e suites cover which services". Consumed by
-# tests/e2e/select-suites (Bead 4) to derive targeted suite selection from
-# a PR diff.
-#
-# See docs/development/improvements/e2e-targeted-suite-selection.md.
-#
-# Editing rules
-# -------------
-# - subscribes / publishes: subjects this service receives / sends. Both
-#   lists use the literal subject string, or a "<prefix>.*" wildcard for
-#   pattern subscriptions (e.g. "ebs.snapshot.*").
-# - depends_on: subjects this service requires from peers. Used by the
-#   selector to compute "if X changes, which suites might break".
-# - Subjects must match what is actually referenced in source. The lint
-#   in Bead 5 greps `"(ec2|vpc|ebs|s3|iam|elbv2|spinifex)\.[a-zA-Z._-]+"`
-#   and fails if any subject is absent from this manifest.
-
-version: 1
-
-services:
-
-  awsgw:
-    path: spinifex/runtime/roles/awsgw
-    # TLS frontend, SigV4 auth. Pure router: HTTP → NATS request/response.
-    # No subscribes; only publishes (via NATSRequest).
-    subscribes: []
-    publishes:
-      - "ec2.*"
-      - "elbv2.*"
-      - "acm.*"
-      - "route53.*"
-      - "iam.*"
-      - "rds.*"
-      - "ecs.*"
-      - "ochre.vector.*"
-    depends_on:
-      - "ec2.*"
-      - "elbv2.*"
-      - "acm.*"
-      - "route53.*"
-      - "iam.*"
-      - "rds.*"
-      - "ecs.*"
-      - "ochre.vector.*"
-
-  spinifex_daemon:
-    path: spinifex/runtime/roles/spinifex
-    # Orchestrator. Owns most ec2.* + all elbv2.* + iam event publication.
-    # Translates VPC / EBS / S3 operations into vpc.* / ebs.* / s3.* and
-    # forwards to vpcd / viperblockd / predastore.
-    #
-    # Source layout: spinifex/runtime/roles/spinifex (entry), spinifex/runtime/daemon
-    # (subscribe registration), spinifex/domains/<domain> and the legacy
-    # spinifex/handlers/{ecs,eks,elbv2,iam,rds} packages (per-subject handler
-    # bodies). Selector treats these paths as belonging to this service via
-    # the `additional_paths` field.
-    additional_paths:
-      - spinifex/runtime/daemon
-      - spinifex/domains/ec2/account
-      - spinifex/domains/ec2/eigw
-      - spinifex/domains/ec2/eip
-      - spinifex/domains/ec2/igw
-      - spinifex/domains/ec2/image
-      - spinifex/domains/ec2/instance
-      - spinifex/domains/ec2/key
-      - spinifex/domains/ec2/launchtemplate
-      - spinifex/domains/ec2/natgw
-      - spinifex/domains/ec2/placementgroup
-      - spinifex/domains/ec2/routetable
-      - spinifex/domains/ec2/snapshot
-      - spinifex/domains/ec2/spotinstance
-      - spinifex/domains/ec2/tags
-      - spinifex/domains/ec2/volume
-      - spinifex/domains/ec2/volumestate
-      - spinifex/domains/ec2/vpc
-      - spinifex/domains/elbv2
-      - spinifex/domains/acm
-      - spinifex/domains/route53
-      - spinifex/domains/iam
-      - spinifex/domains/ecs
-      - spinifex/domains/rds/instance
-      - spinifex/domains/rds/subnetgroup
-      - spinifex/domains/rds/parametergroup
-      - spinifex/domains/rds/engine
-      - spinifex/domains/ecs/taskdefinition
-      - spinifex/domains/eks/cluster
-      - spinifex/domains/eks/access
-      - spinifex/domains/eks/addon
-      # Guest add-on wire contract (subject, report, staged-manifest body).
-      - contracts/eks/v1
-      # ECS agent/scheduler Layer-2 bus wire contract (subjects, message types).
-      - contracts/ecs/v1
-      - spinifex/domains/ochre/vector
-    subscribes:
-      # EC2 instance lifecycle.
-      - "ec2.RunInstances"
-      - "ec2.DescribeInstances"
-      - "ec2.DescribeStoppedInstances"
-      - "ec2.DescribeTerminatedInstances"
-      - "ec2.DescribeInstanceAttribute"
-      - "ec2.DescribeInstanceStatus"
-      - "ec2.DescribeInstanceTypes"
-      - "ec2.ModifyInstanceAttribute"
-      - "ec2.ModifyVolume"
-      - "ec2.start"
-      - "ec2.terminate"
-      - "ec2.launch"
-      # contracts/ec2/v1 owns the command payload and this subject vocabulary.
-      # The wildcard is intentionally one final instance-ID token, not `>`.
-      - "ec2.cmd.*"
-      # EC2 VPC / network ops handled by daemon before delegation to vpcd.
-      - "ec2.CreateVpc"
-      - "ec2.DescribeVpcs"
-      - "ec2.CreateSubnet"
-      - "ec2.CreateInternetGateway"
-      - "ec2.CreateKeyPair"
-      - "ec2.CreateImage"
-      - "ec2.DescribePlacementGroups"
-      # Cluster coordination subjects.
-      - "ec2.ReserveClusterNode"
-      - "ec2.FinalizeClusterInstances"
-      # An EIP association reaches whichever node holds the instance, which is
-      # rarely the node that served the request (fan-out, no queue group).
-      - "ec2.eni.public-ip-changed"
-      # contracts/cluster/v1 owns the node status/capacity payload and subject.
-      - "spinifex.node.status"
-      # contracts/cluster/v1 owns the VM inventory payload and subject.
-      - "spinifex.node.vms"
-      - "spinifex.nodes.discover"
-      - "spinifex.cluster.shutdown.progress"
-      # Coordinated cluster-shutdown phases (fan-out, no queue group).
-      - "spinifex.cluster.shutdown.gate"
-      - "spinifex.cluster.shutdown.drain"
-      - "spinifex.cluster.shutdown.storage"
-      - "spinifex.cluster.shutdown.persist"
-      - "spinifex.cluster.shutdown.infra"
-      # Operator-plane Predastore topology contract (contracts/operator/v1).
-      - "spinifex.storage.config"
-      # ELBv2 control plane (domains/elbv2/service_nats.go topic map).
-      - "elbv2.CreateLoadBalancer"
-      - "elbv2.DeleteLoadBalancer"
-      - "elbv2.DescribeLoadBalancers"
-      - "elbv2.CreateTargetGroup"
-      - "elbv2.DeleteTargetGroup"
-      - "elbv2.DescribeTargetGroups"
-      - "elbv2.RegisterTargets"
-      - "elbv2.DeregisterTargets"
-      - "elbv2.DescribeTargetHealth"
-      - "elbv2.CreateListener"
-      - "elbv2.DeleteListener"
-      - "elbv2.DescribeListeners"
-      # ACM-alike certificate store (domains/acm; subscribed in daemon.go).
-      - "acm.ImportCertificate"
-      - "acm.DescribeCertificate"
-      - "acm.ListCertificates"
-      - "acm.DeleteCertificate"
-      # Route 53 hosted zones (domains/route53; subscribed in daemon.go).
-      - "route53.*"
-      # Operator-only zone activation and delegation, sent by spx admin dns.
-      - "route53.admin.*"
-      - "elbv2.AddListenerCertificates"
-      - "elbv2.RemoveListenerCertificates"
-      - "elbv2.DescribeListenerCertificates"
-      - "elbv2.DescribeSSLPolicies"
-      # RDS control plane (domains/rds/instance/subjects.go). One wildcard rather than
-      # 30 literals: every action subject is registered from a constant, so a
-      # per-subject list would be a hand-mirror of that file with nothing
-      # checking it.
-      - "rds.*"
-      # ECS control plane (domains/ecs/service_nats.go, registered in
-      # daemon.go). One wildcard, as for RDS: every action subject is a literal
-      # in those two files, so a per-subject list would only mirror them.
-      - "ecs.*"
-      # Layer-2 agent bus: register and health are wildcard subscriptions over
-      # {account}.{db-instance}, so the subject is built at runtime and only the
-      # prefix is declarable.
-      - "rds.bus.*"
-      # ECS Layer-2 scheduler/agent bus (contracts/ecs/v1): account, cluster and
-      # instance/task IDs are runtime-built, so only the prefix is declarable.
-      # Only the leader holds these subscriptions, so there is no queue group.
-      - "ecs.bus.*"
-      # Ochre vector store tenant surface (domains/ochre/vector/api.go),
-      # gated behind config.OchreVectorConfig.Enabled (default off).
-      - "ochre.vector.createIndex"
-      - "ochre.vector.deleteIndex"
-      - "ochre.vector.listIndexes"
-      - "ochre.vector.ingest"
-      - "ochre.vector.describeJob"
-      - "ochre.vector.query"
-      - "ochre.vector.listJobs"
-      - "ochre.vector.stopJob"
-      # Fence route owned by contracts/viperblockd/legacy/v1: viperblockd
-      # announces a torn-down export so this node stops the guest that was
-      # using it. "ebs.fenced" is the single-node form; a named node is
-      # addressed as "ebs.<node>.fenced".
-      - "ebs.fenced"
-    publishes:
-      # → vpcd
-      - "vpc.create"
-      - "vpc.delete"
-      - "vpc.create-subnet"
-      - "vpc.delete-subnet"
-      - "vpc.create-sg"
-      - "vpc.delete-sg"
-      - "vpc.update-sg"
-      - "vpc.create-port"
-      - "vpc.delete-port"
-      - "vpc.update-port-sgs"
-      # EC2-to-network projection owned by contracts/ec2/v1. The network
-      # subscriber realizes it in OVN without reading EC2 private state.
-      - "vpc.igw-attach"
-      - "vpc.igw-detach"
-      - "vpc.add-nat-gateway"
-      - "vpc.delete-nat-gateway"
-      - "vpc.add-nat"
-      - "vpc.delete-nat"
-      - "vpc.dhcp.acquire"
-      - "vpc.dhcp.release"
-      # Per-subnet route-table projections owned by domains/ec2/routetable:
-      # IGW default-route reachability and the egress gate/ungate decision it
-      # implies, published fire-and-forget on every route-table mutation.
-      - "vpc.add-igw-route"
-      - "vpc.delete-igw-route"
-      - "vpc.gate-subnet-egress"
-      - "vpc.ungate-subnet-egress"
-      # Egress-only reroute removal for a single system instance
-      # (domains/eks/cluster). Fire-and-forget. vpc.add-system-egress has no
-      # current producer.
-      - "vpc.delete-system-egress"
-      # → viperblockd
-      # The migrated legacy ebs.config and ebs.sync payloads and routes are
-      # owned by contracts/viperblockd/legacy/v1. They are transitional and
-      # distinct from the generic providers/ebs protocol; other legacy ebs.*
-      # routes remain to be migrated separately.
-      - "ebs.mount"
-      - "ebs.unmount"
-      - "ebs.delete"
-      - "ebs.sync"
-      - "ebs.snapshot.*"
-      - "ebs.config"
-      - "ebs.config.*"
-      # Abandon route owned by contracts/viperblockd/legacy/v1: the daemon
-      # asks viperblockd to tear an export down without sealing. "ebs.abandon"
-      # is the single-node form; a named node addresses "ebs.<node>.abandon".
-      - "ebs.abandon"
-      # → predastore (S3 / AMI / keypair store)
-      - "s3.*"
-      # → every daemon, so the one holding the instance reports the new address
-      - "ec2.eni.public-ip-changed"
-      # IAM lifecycle event
-      - "iam.account.created"
-      # → the in-guest rds-agent: set-password and the like, published on the
-      # per-instance command subject and answered on its reply subject.
-      - "rds.bus.*"
-      # ECS service-reconcile self-wake (contracts/ecs/v1): the worker that
-      # recorded a task-state change is rarely the leader, so this fans the
-      # wake out and the leader's own "ecs.bus.*" subscription picks it up.
-      - "ecs.bus.*"
-    depends_on:
-      - "vpc.*"
-      - "ebs.*"
-      - "s3.*"
-
-  vpcd:
-    path: spinifex/runtime/roles/vpcd
-    # OVN-backed VPC / subnet / SG / NAT / IGW data plane.
-    subscribes:
-      - "vpc.create"
-      - "vpc.delete"
-      - "vpc.create-subnet"
-      - "vpc.delete-subnet"
-      - "vpc.create-sg"
-      - "vpc.delete-sg"
-      - "vpc.update-sg"
-      - "vpc.create-port"
-      - "vpc.delete-port"
-      - "vpc.update-port-sgs"
-      - "vpc.igw-attach"
-      - "vpc.igw-detach"
-      - "vpc.add-nat-gateway"
-      - "vpc.delete-nat-gateway"
-      - "vpc.add-nat"
-      - "vpc.delete-nat"
-      - "vpc.dhcp.acquire"
-      - "vpc.dhcp.release"
-      - "vpc.add-igw-route"
-      - "vpc.delete-igw-route"
-      - "vpc.gate-subnet-egress"
-      - "vpc.ungate-subnet-egress"
-      - "vpc.add-system-egress"
-      - "vpc.delete-system-egress"
-    publishes:
-      - "vpc.port-status"
-      - "vpc.dhcp.lease-changed"
-      - "vpc.dhcp.owner-check"
-    depends_on: []
-
-  viperblockd:
-    path: spinifex/runtime/roles/viperblockd
-    # EBS data plane. Backed by the viperblock storage engine in the
-    # mulga/viperblock submodule. Submodule pointer bump → selector forces
-    # "run all" (contract changes can be invisible at the spinifex source
-    # level).
-    # The migrated legacy ebs.config and ebs.sync routes are versioned by
-    # contracts/viperblockd/legacy/v1, not by providers/ebs. Other legacy
-    # ebs.* routes remain to be migrated separately.
-    # The dirty-marker JetStream KV bucket VIPERBLOCK_VOLUME_DIRTY is owned
-    # by the same contract; not a NATS subject, so not listed below.
-    #
-    # Source layout: spinifex/runtime/roles/viperblockd (role: process lifecycle,
-    # NATS subscription wiring) and spinifex/providers/ebs/viperblock (the
-    # Viperblock EBS provider adapter the role wires up). Selector treats the
-    # adapter path as belonging to this service via `additional_paths`.
-    additional_paths:
-      - spinifex/providers/ebs/viperblock
-    subscribes:
-      - "ebs.mount"
-      - "ebs.unmount"
-      - "ebs.delete"
-      - "ebs.sync"
-      - "ebs.snapshot.*"
-      - "ebs.config"
-      - "ebs.config.*"
-      # Abandon route (contracts/viperblockd/legacy/v1): request/reply from
-      # the daemon, answered by the Viperblock adapter on the named node.
-      - "ebs.abandon"
-    publishes:
-      - "ebs.mount.response"
-      # Fence route (contracts/viperblockd/legacy/v1): announced after the
-      # Viperblock adapter tears an export down because its lease moved.
-      - "ebs.fenced"
-      - "ebs.unmount.response"
-    depends_on: []
-
-  predastore:
-    path: spinifex/runtime/roles/predastore
-    # S3 / AMI / keypair storage. External submodule under mulga/predastore;
-    # this dir contains only the spinifex-side adapter. Submodule pointer
-    # bump → selector forces "run all".
-    subscribes:
-      - "s3.*"
-    publishes: []
-    depends_on: []
-
-  nats:
-    path: spinifex/runtime/roles/nats
-    # NATS broker bootstrap / cluster config. Not a NATS pub/sub
-    # participant. Any change here is infra-path; selector forces
-    # "run all".
-    subscribes: []
-    publishes: []
-    depends_on: []
-
-  spinifexui:
-    path: spinifex/runtime/roles/spinifexui
-    # Web UI. Talks to awsgw over HTTP, never NATS. Diffs limited to this
-    # dir do not need any e2e suite to run.
-    subscribes: []
-    publishes: []
-    depends_on: []
-    covers_no_suite: true
-
-
-# Fixture → service mapping. Each Ensure* helper in
-# tests/e2e/harness/fixtures.go exercises this set of services; the
-# Bead 4 selector AST-scans test files for harness.Ensure* call sites
-# and resolves coverage via this block.
-fixtures:
-
-  EnsureKeyPair:
-    services: [spinifex_daemon, predastore]
-
-  EnsureAMI:
-    services: [spinifex_daemon, predastore, viperblockd]
-
-  EnsureDefaultVPC:
-    services: [vpcd, spinifex_daemon]
-
-  EnsureSubnet:
-    services: [vpcd, spinifex_daemon]
-
-  EnsureSG:
-    services: [vpcd, spinifex_daemon]
-
-  EnsureInstance:
-    services: [spinifex_daemon, viperblockd, vpcd]
-
-  EnsureVolume:
-    services: [spinifex_daemon, viperblockd]
-
-  EnsureSnapshot:
-    services: [spinifex_daemon, viperblockd, predastore]
-
-  EnsureNATGateway:
-    services: [vpcd, spinifex_daemon]
-
-  EnsureLoadBalancer:
-    services: [spinifex_daemon, awsgw]
-
-
-suites:
-
-  e2e-cert:
-    path: tests/e2e/cert
-    covers: [awsgw, spinifex_daemon, vpcd, predastore]
-
-  e2e-lb:
-    path: tests/e2e/lb
-    covers: [awsgw, spinifex_daemon, vpcd, viperblockd, predastore]
-    # Exercises full ELBv2 control plane (domains/elbv2) + data plane
-    # (LB agent on the system VM). ELBv2 is part of spinifex_daemon in
-    # this manifest; revisit if it splits into a standalone service.
-
-  e2e-single:
-    path: tests/e2e/single
-    covers: [awsgw, spinifex_daemon, vpcd, viperblockd, predastore]
-    # Full single-node lifecycle: VPC, AMI, keypair, instance, volume,
-    # snapshot, IAM. Currently runs all phases sequentially under one
-    # `TestSingleNode`. Bead 3 (single-fixture-migration) splits this
-    # into per-test fixture-driven coverage so the selector can pick
-    # subtests, not just whole suites.
-
-  e2e-iam:
-    path: tests/e2e/iam
-    covers: [awsgw, spinifex_daemon, vpcd, viperblockd, predastore]
-    # "Everything IAM": the IAM/STS control plane plus the instance-identity
-    # tests (instance-profile association + IMDS), which boot guest VMs — OVN
-    # via vpcd, AMI via predastore, root volume via viperblockd, lifecycle via
-    # spinifex_daemon. Single-mode, VM-booting leg.
-
-  e2e-ecr:
-    path: tests/e2e/ecr
-    covers: [awsgw, spinifex_daemon, predastore]
-    # ECR control plane (domains/ecr, gateway/ecr*) + OCI data plane: blobs and
-    # manifests in predastore, control + auth bridge in awsgw, KV via
-    # spinifex_daemon. docker/crane/skopeo data-plane subtests skip when the
-    # client binary or registry-host resolution is absent.
-
-  e2e-bedrock:
-    path: tests/e2e/bedrock
-    covers: [awsgw]
-    # Ochre (AWS-Bedrock) control plane: static catalog list/get served
-    # straight from awsgw (gateway/bedrock.go, domains/ochre/) — no NATS
-    # hop to spinifex_daemon. Converse (Anthropic-direct / self-host vLLM)
-    # subtests are gated on backend availability and skip by default.
-
-  e2e-ecs:
-    path: tests/e2e/ecs
-    covers: [awsgw, spinifex_daemon, vpcd, viperblockd, predastore]
-    # ECS data plane end-to-end: cluster + a real container instance launched
-    # from the spinifex-ecs-node AMI (boots, registers over the gateway, runs
-    # tasks through containerd), standalone tasks in awsvpc/bridge/host modes,
-    # and an awsvpc service fronted by an ELBv2 target group. Boots a guest VM,
-    # so vpcd (OVN + awsvpc ENI), viperblockd (root volume), predastore (AMI)
-    # are all in the path.
-
-  e2e-rds:
-    path: tests/e2e/rds
-    covers: [awsgw, spinifex_daemon, vpcd, viperblockd, predastore]
-    # RDS create/describe/connect: CreateDBInstance boots a dual-NIC DB VM from
-    # the engine AMI (predastore) with a Viperblock data volume (viperblockd) and
-    # a customer ENI (vpcd), and the reconciler flips it to available on the
-    # in-guest agent's first healthy heartbeat over awsgw. Every test deletes the
-    # instances it creates, so the suite needs no gate beyond SPINIFEX_E2E.
-
-  e2e-ddil:
-    path: tests/e2e/ddil
-    covers: [awsgw, spinifex_daemon, vpcd, viperblockd, predastore]
-    # DDIL = degraded / disconnected / intermittent / low-bandwidth.
-    # Exercises the full daemon stack under network failure injection;
-    # any service touching the wire path is in scope.
-
-  e2e-route53:
-    path: tests/e2e/route53
-    covers: [awsgw, spinifex_daemon, predastore]
-    # Hosted zones end to end: delegation check, publish to the northstar zone
-    # store in predastore, and every node answering before a change is INSYNC.
-
-  # Suites below are bash-only today. Entries activate as the Go ports
-  # land (Bead 2 fixtures unblock new-shape ports for multinode etc.).
-  #
-  # e2e-multinode:
-  #   path: spinifex/tests/e2e/multinode
-  #   covers: [awsgw, spinifex_daemon, vpcd, viperblockd, predastore]
-  # e2e-reboot:
-  #   path: spinifex/tests/e2e/reboot
-  #   covers: [spinifex_daemon, viperblockd]
-  # e2e-tofu:
-  #   path: spinifex/tests/e2e/tofu
-  #   covers: [awsgw, spinifex_daemon, vpcd]
-  # e2e-bm:
-  #   path: spinifex/tests/e2e/baremetal
-  #   covers: [awsgw, spinifex_daemon, vpcd, viperblockd, predastore]
-  #
-  # e2e-natuplink activates once CI can provision an external_mode=nat
-  # node (setup-ovn.sh --nat-uplink + spx admin init --external-mode=nat).
-  # e2e-natuplink:
-  #   path: spinifex/tests/e2e/natuplink
-  #   covers: [awsgw, spinifex_daemon, vpcd, viperblockd, predastore]
-
-
-# Special-case rules consumed by the selector (Bead 4). Hard-coded in
-# the tool, listed here for visibility.
-#
-# infra_paths_force_all:
-#   - spinifex/runtime/roles/nats/**
-#   - spinifex/tests/e2e/harness/**
-#   - spinifex/tests/e2e/lib/**
-#   - spinifex/.github/workflows/e2e-go.yml
-#   - spinifex/.github/scripts/**
-#   - spinifex/Makefile
-#   - mulga/scripts/tofu-cluster/**
-#
-# submodule_force_all:
-#   - mulga/predastore (submodule pointer)
-#   - mulga/viperblock (submodule pointer)
-#
-# many_services_force_all_threshold: 3   # ≥ N services touched → run all
@@ -69,8 +69,8 @@ Clone the Terraform examples from the Spinifex repository:
 ```bash
 git clone --depth 1 --filter=blob:none --sparse https://github.com/mulgadc/spinifex.git spinifex-tf
 cd spinifex-tf
-git sparse-checkout set docs/terraform
-cd docs/terraform/bastion-private-subnet
+git sparse-checkout set docs/terraform-workbooks
+cd docs/terraform-workbooks/bastion-private-subnet
 ```

 Or create a `main.tf` file and paste the full configuration below.
@@ -1,7 +1,7 @@
 ---
 title: "Nginx Web Server (Load Balanced)"
 seoTitle: "Load-Balanced Nginx Behind an AWS ALB — Spinifex Docs"
-description: "Deploy a VPC with two private EC2 instances running Nginx behind an internet-facing Application Load Balancer on Spinifex, using Terraform or OpenTofu."
+description: "Deploy Nginx EC2 instances behind an internet-facing Application Load Balancer on Spinifex with Terraform, then spread them across nodes and test failover."
 category: "Terraform Workbooks"
 tags:
  - terraform
@@ -10,6 +10,8 @@ tags:
  - elbv2
  - alb
  - vpc
+  - placement-group
+  - resiliency
  - workbook
 resources:
  - title: "Terraform AWS Provider"
@@ -22,7 +24,7 @@ resources:

 # Terraform: Nginx Web Servers with ALB

-> Deploys two Nginx web server instances behind an Application Load Balancer on Spinifex.
+> Deploys Nginx web server instances behind an Application Load Balancer on Spinifex, optionally spread one per node so the site survives losing a node.

 ## Table of Contents

@@ -34,18 +36,25 @@ resources:

 ## Overview

-Deploy two Nginx web servers behind an internet-facing Application Load Balancer on Spinifex using Terraform/OpenTofu. This workbook provisions a VPC with public and private subnets, an internet gateway and NAT Gateway, route tables, security group, SSH key pair, an application load balancer (ALB) and two EC2 instances with cloud-init user-data that installs and starts Nginx. Only the ALB is reachable from outside the VPC — the Nginx instances live in the **private subnets** and reach the internet only for cloud-init bootstrapping via the NAT Gateway.
+Deploy Nginx web servers behind an internet-facing Application Load Balancer on Spinifex using Terraform or OpenTofu.
+This workbook provisions a VPC with two public subnets, an internet gateway, a route table, a security group, an SSH key pair, an Application Load Balancer (ALB) and two or more EC2 instances with cloud-init user-data that installs and starts Nginx.
+The web servers get public IPs so cloud-init can install Nginx, and the ALB reaches them over the VPC by their private addresses.
+
+On a cluster of three or more nodes, a spread placement group puts each web server on a different node.
+Losing a node then takes out at most one server, the ALB stops sending it traffic, and Spinifex restarts the lost instances on the surviving nodes.
+[Step 6](#step-6-test-resiliency) shows how to watch that happen.

 <p align="center">
-  <img src="../../../.github/assets/diagrams/tf-nginx-alb.svg" alt="Nginx + ALB VPC — IGW, two public subnets with ALB ENIs and NAT GW, two private subnets hosting nginx" width="900">
+  <img src="../../../.github/assets/diagrams/tf-nginx-alb.svg" alt="Nginx + ALB VPC — IGW, two public subnets with the ALB and the nginx web servers" width="900">
 </p>

 **What you'll learn:**

 - Configuring the AWS Terraform provider to target Spinifex
- Creating a VPC with public + private subnets, an IGW and a NAT Gateway
- Provisioning a multi-AZ internet-facing ALB fronting private workers
- Provisioning an EC2 instance with cloud-init user-data
+- Creating a VPC with public subnets and an internet gateway
+- Provisioning an internet-facing ALB in front of a group of web servers
+- Provisioning EC2 instances with cloud-init user-data
+- Spreading instances across nodes with a placement group
 - Generating SSH key pairs with the TLS provider

 **What gets created**
@@ -53,22 +62,33 @@ Deploy two Nginx web servers behind an internet-facing Application Load Balancer
 | Resource | Name | Purpose |
 |---|---|---|
 | VPC | `nginx-alb-vpc` | Isolated network (10.20.0.0/16) |
-| Public Subnets | `nginx-alb-public-a`, `nginx-alb-public-b` | Two AZs hosting the ALB and NAT Gateway |
-| Private Subnets | `nginx-alb-private-a`, `nginx-alb-private-b` | Two AZs hosting the Nginx workers |
+| Public Subnets | `nginx-alb-public-a`, `nginx-alb-public-b` | Host the ALB and the web servers |
 | Internet Gateway | `nginx-alb-igw` | Routes internet traffic for the public subnets |
-| Elastic IP | `nginx-alb-nat-eip` | Public address for the NAT Gateway |
-| NAT Gateway | `nginx-alb-nat` | Outbound internet for the private subnets (cloud-init apt bootstrap) |
 | Security Group | `nginx-alb-sg` | Allows SSH (22) and HTTP (80) inbound |
-| EC2 Instances | `nginx-alb-1`, `nginx-alb-2` | Ubuntu 26.04 with Nginx via cloud-init (private subnets) |
+| Placement Group | `nginx-alb-spread` | Only with `placement_strategy`; spread puts one web server per node |
+| EC2 Instances | `nginx-alb-1`, `nginx-alb-2`, … | Ubuntu with Nginx via cloud-init, one per `server_count` |
 | ALB | `nginx-alb` | Internet-facing Application Load Balancer on port 80 |
-| Target Group | `nginx-alb-tg` | HTTP health-checked group for both instances |
+| Target Group | `nginx-alb-tg` | HTTP health-checked group of every web server |
 | Listener | HTTP :80 | Forwards traffic to the target group |

+**Variables**
+
+| Variable | Default | Purpose |
+|---|---|---|
+| `instance_type` | `t3.small` | Instance type of the web servers |
+| `server_count` | `2` | Number of web servers behind the ALB |
+| `placement_strategy` | none | `spread` puts each web server on a different node; `cluster` puts them all on one |
+| `region` | `ap-southeast-2` | Spinifex region |
+| `spinifex_endpoint` | `https://127.0.0.1:9999` | Spinifex AWS endpoint |
+
+A spread placement group on Spinifex spreads across nodes, so `server_count` can be at most the number of nodes in the cluster.
+
 **Prerequisites:**

 - Spinifex installed and running (see [Installing Spinifex](/docs/install))
- An Ubuntu 26.04 AMI imported (see [Setting Up Your Cluster](/docs/setting-up-your-cluster))
+- An Ubuntu 26.04 or 24.04 AMI imported (see [Setting Up Your Cluster](/docs/setting-up-your-cluster))
 - OpenTofu or Terraform installed
+- For [Step 6](#step-6-test-resiliency), a cluster of three nodes

 ## Instructions

@@ -79,20 +99,20 @@ Clone the Terraform examples from the Spinifex repository:
 ```bash
 git clone --depth 1 --filter=blob:none --sparse https://github.com/mulgadc/spinifex.git spinifex-tf
 cd spinifex-tf
-git sparse-checkout set docs/terraform
-cd docs/terraform/nginx-alb
+git sparse-checkout set docs/terraform-workbooks
+cd docs/terraform-workbooks/nginx-alb
 ```

 Or create a `main.tf` file and paste the full configuration below.

 <!-- INCLUDE: main.tf lang:hcl -->

-### Step 2. Install Debian AMI
+### Step 2. Import the Ubuntu AMI

-Install the Debian 13 AMI which is used in the example to host the `nginx` webservers as an EC2 instance.
+Import the Ubuntu 26.04 AMI the web servers run on, from any node of the cluster:

 ```bash
-spx admin images import --name debian-13-x86_64
+spx admin images import --name ubuntu-26.04-x86_64
 ```

 > **Note:** The load balancer itself runs as a direct-boot QEMU microvm using the kernel + initramfs bundled with the Spinifex distribution (`/usr/share/spinifex/microvm/`), so no separate LB AMI import is required.
@@ -133,9 +153,16 @@ Next, apply and launch the template:
 tofu apply
 ```

+On a three-node cluster, launch three web servers instead, one on each node:
+
+```bash
+tofu apply -var server_count=3 -var placement_strategy=spread
+```
+
 ### Step 5. Verify

-> **Note:** EC2 instances can take 30+ seconds to boot after apply, and the NAT Gateway must be `available` before cloud-init on the workers can reach the apt repository. If the ALB returns 5xx or HTTP is unreachable, wait and retry — the target group health checks need a moment to mark both instances healthy once Nginx has installed.
+> **Note:** EC2 instances can take 30+ seconds to boot after apply, and Nginx installs after that.
+> If the ALB returns 5xx or HTTP is unreachable, wait and retry, since the target group health checks need a moment to mark the instances healthy once Nginx has installed.

 The ALB is internet-facing, but the DNS name Spinifex returns (`*.elb.spinifex.local`) will not resolve from your host. Fetch the ALB's public IP with the AWS CLI:

@@ -145,7 +172,7 @@ ALB_IP=$(aws elbv2 describe-load-balancers --names nginx-alb \
  --output text)
 ```

-Then hit the ALB — successive requests should alternate between Server 1 and Server 2:
+Then hit the ALB, and successive requests rotate between the servers:

 ```bash
 curl http://$ALB_IP
@@ -154,7 +181,8 @@ curl http://$ALB_IP

 Open `http://$ALB_IP` in your browser and refresh to see the page alternate content served from each instance.

-The Nginx instances themselves only have private IPs (see the `instance_1_private_ip` / `instance_2_private_ip` outputs) and are only reachable from inside the VPC — go through the ALB.
+Each web server also has a public IP of its own, listed by `tofu output instance_public_ips`.
+The ALB reaches them by their private IPs, listed by `tofu output instance_private_ips`.

 Check target health via AWS CLI:

@@ -166,12 +194,40 @@ TG_ARN=$(aws elbv2 describe-target-groups \
 aws elbv2 describe-target-health --target-group-arn $TG_ARN
 ```

+### Step 6. Test resiliency
+
+On a cluster of three nodes, apply with `-var server_count=3 -var placement_strategy=spread`, then check that each web server is on its own node:
+
+```bash
+spx get vms
+```
+
+The `NODE` column names each node once for the three `t3.small` web servers, and the load balancer is the `sys.micro` instance.
+
+Request the page once a second in a second terminal:
+
+```bash
+while :; do
+  printf '%s %s\n' "$(date +%T)" "$(curl -s -m2 "http://$ALB_IP/" | grep -o 'Server [0-9]' || echo FAILED)"
+  sleep 1
+done
+```
+
+Then power off one node, the way a hardware failure would, without shutting it down first.
+On Vultr, `deploy-vultr.sh stop node3` does this, as described in [Running Spinifex on Vultr](/docs/vultr-integration#8-test-resiliency).
+
+Within about a minute the load balancer stops sending requests to the lost web server, and the other two answer every request.
+About three minutes after the power-off, Spinifex has restarted the lost web server on a surviving node with the same instance ID, private IP and public IP, and the load balancer sends it requests again once its health checks pass.
+Power the node back on and wait until `spx get nodes` shows it `Ready` before you test another node.
+
 ### Cleanup

 ```bash
 tofu destroy
 ```

+If you applied with `-var` options, pass the same ones to `tofu destroy`.
+
 ## Troubleshooting

 ### AMI Not Found
@@ -213,18 +269,19 @@ If targets stay unhealthy, verify the instances are running:
 aws ec2 describe-instances --profile spinifex
 ```

-If cloud-init on the workers never finished, confirm the NAT Gateway is `available` (the private subnets rely on it for outbound apt access):
-
-```bash
-aws ec2 describe-nat-gateways --query 'NatGateways[].[NatGatewayId,State]'
-```
+If cloud-init on the workers never finished, SSH to one by its public IP and read its log, as described in [Nginx Not Responding](#nginx-not-responding).

 ### Nginx Not Responding

-The Nginx instances have no public IP, so you can't SSH in directly from your host. If you need to inspect cloud-init logs, launch a small jump host in the same VPC or run commands via the Spinifex console, then:
+SSH to the web server by its public IP, from `tofu output instance_public_ips`, and check cloud-init and Nginx:

 ```bash
-ssh -i nginx-alb-demo.pem ubuntu@<instance_private_ip>
+ssh -i nginx-alb-demo.pem ubuntu@<instance_public_ip>
 sudo journalctl -u cloud-init --no-pager
 sudo systemctl status nginx
 ```
+
+### Spread Launch Fails With InsufficientInstanceCapacity
+
+A spread placement group puts at most one of its instances on each node.
+Set `server_count` no higher than the number of nodes, or leave `placement_strategy` empty.
@@ -12,10 +12,16 @@
 #
 # Demonstrates: VPC, public subnets, internet gateway, route tables,
 # security group, key pair, cloud-init user-data, EC2 instances with
-# auto-assigned public IPs, ALB, target group, and listener.
+# auto-assigned public IPs, an optional placement group, ALB, target group,
+# and listener.
+#
+# server_count sets how many web servers launch (default 2). Setting
+# placement_strategy = "spread" puts each one on a different Spinifex node,
+# so a three-node cluster with server_count = 3 can lose any node and keep
+# serving through the ALB.
 #
 # Usage:
-#   cd spinifex/docs/terraform/nginx-alb
+#   cd spinifex/docs/terraform-workbooks/nginx-alb
 #   export AWS_PROFILE=spinifex
 #   tofu init && tofu apply
 #
@@ -28,6 +34,9 @@
 #
 # Then:
 #   curl http://<alb_public_ip>    # Load-balanced Nginx (alternates between instances)
+#
+# Three servers, one per node:
+#   tofu apply -var server_count=3 -var placement_strategy=spread

 terraform {
  required_version = ">= 1.6.0"
@@ -62,6 +71,28 @@ variable "instance_type" {
  default = "t3.small"
 }

+variable "server_count" {
+  type        = number
+  default     = 2
+  description = "Number of Nginx web servers behind the ALB"
+
+  validation {
+    condition     = var.server_count >= 1
+    error_message = "server_count must be at least 1."
+  }
+}
+
+variable "placement_strategy" {
+  type        = string
+  default     = ""
+  description = "Placement group strategy for the web servers: spread (one per node), cluster (all on one node), or empty for none"
+
+  validation {
+    condition     = contains(["", "spread", "cluster"], var.placement_strategy)
+    error_message = "placement_strategy must be spread, cluster, or empty."
+  }
+}
+
 variable "spinifex_endpoint" {
  type        = string
  default     = "https://127.0.0.1:9999"
@@ -241,14 +272,28 @@ resource "aws_security_group" "web" {
 }

 # ---------------------------------------------------------------------------
-# EC2 Instances — two Nginx servers with distinct landing pages
+# Placement Group — optional; spread puts each web server on its own node
 # ---------------------------------------------------------------------------

-resource "aws_instance" "nginx_1" {
-  ami           = data.aws_ami.ubuntu.id
-  instance_type = var.instance_type
+resource "aws_placement_group" "web" {
+  count    = var.placement_strategy == "" ? 0 : 1
+  name     = "nginx-alb-${var.placement_strategy}"
+  strategy = var.placement_strategy
+}

-  subnet_id              = aws_subnet.public_a.id
+# ---------------------------------------------------------------------------
+# EC2 Instances — Nginx servers with distinct landing pages, alternating
+# between the two public subnets
+# ---------------------------------------------------------------------------
+
+resource "aws_instance" "nginx" {
+  count = var.server_count
+
+  ami             = data.aws_ami.ubuntu.id
+  instance_type   = var.instance_type
+  placement_group = one(aws_placement_group.web[*].name)
+
+  subnet_id              = count.index % 2 == 0 ? aws_subnet.public_a.id : aws_subnet.public_b.id
  vpc_security_group_ids = [aws_security_group.web.id]
  key_name               = aws_key_pair.nginx.key_name

@@ -266,7 +311,7 @@ resource "aws_instance" "nginx_1" {
    <head><title>Spinifex ALB Demo</title></head>
    <body style="font-family: sans-serif; max-width: 600px; margin: 80px auto;">
      <h1>Hello from Spinifex!</h1>
-      <p><strong>Instance:</strong> $INSTANCE_ID (Server 1)</p>
+      <p><strong>Instance:</strong> $INSTANCE_ID (Server ${count.index + 1})</p>
      <p>This Nginx server is behind an Application Load Balancer.</p>
      <hr>
      <p><small>Provisioned via cloud-init user-data.</small></p>
@@ -280,48 +325,19 @@ resource "aws_instance" "nginx_1" {
  )

  tags = {
-    Name = "nginx-alb-1"
+    Name = "nginx-alb-${count.index + 1}"
  }
 }

-resource "aws_instance" "nginx_2" {
-  ami           = data.aws_ami.ubuntu.id
-  instance_type = var.instance_type
+# Earlier versions declared the two servers separately.
+moved {
+  from = aws_instance.nginx_1
+  to   = aws_instance.nginx[0]
+}

-  subnet_id              = aws_subnet.public_b.id
-  vpc_security_group_ids = [aws_security_group.web.id]
-  key_name               = aws_key_pair.nginx.key_name
-
-  user_data_base64 = base64encode(<<-USERDATA
-    #!/bin/bash
-    set -euo pipefail
-
-    apt-get update -y
-    apt-get install -y nginx
-
-    INSTANCE_ID=$(cat /var/lib/cloud/data/instance-id 2>/dev/null || hostname)
-    cat > /var/www/html/index.html <<HTML
-    <!DOCTYPE html>
-    <html>
-    <head><title>Spinifex ALB Demo</title></head>
-    <body style="font-family: sans-serif; max-width: 600px; margin: 80px auto;">
-      <h1>Hello from Spinifex!</h1>
-      <p><strong>Instance:</strong> $INSTANCE_ID (Server 2)</p>
-      <p>This Nginx server is behind an Application Load Balancer.</p>
-      <hr>
-      <p><small>Provisioned via cloud-init user-data.</small></p>
-    </body>
-    </html>
-    HTML
-
-    systemctl enable nginx
-    systemctl restart nginx
-  USERDATA
-  )
-
-  tags = {
-    Name = "nginx-alb-2"
-  }
+moved {
+  from = aws_instance.nginx_2
+  to   = aws_instance.nginx[1]
 }

 # ---------------------------------------------------------------------------
@@ -365,19 +381,25 @@ resource "aws_lb_target_group" "nginx" {
 }

 # ---------------------------------------------------------------------------
-# Register both instances as targets
+# Register every web server as a target
 # ---------------------------------------------------------------------------

-resource "aws_lb_target_group_attachment" "nginx_1" {
+resource "aws_lb_target_group_attachment" "nginx" {
+  count = var.server_count
+
  target_group_arn = aws_lb_target_group.nginx.arn
-  target_id        = aws_instance.nginx_1.id
+  target_id        = aws_instance.nginx[count.index].id
  port             = 80
 }

-resource "aws_lb_target_group_attachment" "nginx_2" {
-  target_group_arn = aws_lb_target_group.nginx.arn
-  target_id        = aws_instance.nginx_2.id
-  port             = 80
+moved {
+  from = aws_lb_target_group_attachment.nginx_1
+  to   = aws_lb_target_group_attachment.nginx[0]
+}
+
+moved {
+  from = aws_lb_target_group_attachment.nginx_2
+  to   = aws_lb_target_group_attachment.nginx[1]
 }

 # ---------------------------------------------------------------------------
@@ -401,12 +423,11 @@ resource "aws_lb_listener" "http" {

 output "note" {
  value = <<-EOT
-    EC2 instances can take 30+ seconds to boot after apply — if HTTP is
-    unreachable, wait and retry.
+    EC2 instances can take 30+ seconds to boot after apply, and Nginx installs
+    after that — if HTTP is unreachable, wait and retry.

-    The Nginx instances have private IPs only. The ALB DNS name ends in
-    .elb.spinifex.local and will not resolve from your host, so fetch the
-    ALB's public IP with:
+    The ALB DNS name ends in .elb.spinifex.local and will not resolve from
+    your host, so fetch the ALB's public IP with:

      aws elbv2 describe-load-balancers --names nginx-alb \
        --query 'LoadBalancers[0].AvailabilityZones[].LoadBalancerAddresses[].IpAddress' \
@@ -428,18 +449,18 @@ output "alb_dns_name" {
  value = aws_lb.web.dns_name
 }

-output "instance_1_id" {
-  value = aws_instance.nginx_1.id
+output "instance_ids" {
+  value = aws_instance.nginx[*].id
 }

-output "instance_1_private_ip" {
-  value = aws_instance.nginx_1.private_ip
+output "instance_private_ips" {
+  value = aws_instance.nginx[*].private_ip
 }

-output "instance_2_id" {
-  value = aws_instance.nginx_2.id
+output "instance_public_ips" {
+  value = aws_instance.nginx[*].public_ip
 }

-output "instance_2_private_ip" {
-  value = aws_instance.nginx_2.private_ip
+output "placement_group" {
+  value = one(aws_placement_group.web[*].name)
 }
@@ -56,8 +56,8 @@ Clone the Terraform examples from the Spinifex repository:
 ```bash
 git clone --depth 1 --filter=blob:none --sparse https://github.com/mulgadc/spinifex.git spinifex-tf
 cd spinifex-tf
-git sparse-checkout set docs/terraform
-cd docs/terraform/nginx-webserver
+git sparse-checkout set docs/terraform-workbooks
+cd docs/terraform-workbooks/nginx-webserver
 ```

 Or create a `main.tf` file and paste the full configuration below.
@@ -74,8 +74,8 @@ Clone the Terraform examples from the Spinifex repository:
 ```bash
 git clone --depth 1 --filter=blob:none --sparse https://github.com/mulgadc/spinifex.git spinifex-tf
 cd spinifex-tf
-git sparse-checkout set docs/terraform
-cd docs/terraform/s3-webapp
+git sparse-checkout set docs/terraform-workbooks
+cd docs/terraform-workbooks/s3-webapp
 ```

 Or create the files manually and paste the full configuration below.
@@ -30,17 +30,19 @@ The largest suite, 27 top-level tests. Notable entry points:
 - **`TestIMDSDatapathSurvivesVswitchdRestart`** restarts `ovs-vswitchd` under a guest with an Elastic IP, once through systemd and once with SIGKILL. After each, the guest must answer over its EIP and fetch its instance ID through IMDSv2 within 60 seconds, and `br-imds` must hold all 7 of the guest's own flows. It is sequential because it takes the whole node's datapath down.
 - `TestStopStart`, `TestEarlyRebootLiveness`, `TestGuestChurnDurability`, `TestSnapshotLifecycle`, `TestSnapshotBackedLaunch`, `TestCreateImage`, `TestVolumeLifecycle`, `TestSpotInstanceLifecycle`, `TestLaunchTemplateBoot`.
 - `TestNegativeErrorPaths` — the `8a`–`8t` AWS error-shape cases (malformed AMI ID, invalid instance type, volume in use, detach-root forbidden, unknown or malformed instance IDs on the lifecycle verbs, and so on).
- **`TestKVReplication`** — the single-node half of the KV replication policy, and it is not the multi-node assertion with a smaller number in it. On one node one replica per bucket is the correct answer rather than the defect, so what it proves is the opposite: that nothing declared a replica count this deployment cannot place, that `spx admin kv replicas` exits 0 and reports `WANT 1`, and that `--repair` run twice against a cluster with nothing to repair changes no bucket's replica count and leaves every bucket serving. The multi-node suites cannot cover any of that, because on three nodes the unplaceable count is the right one. It lives here rather than in its own suite so it runs in every single-node nightly cell, across all three host and network permutations.
+- **`TestKVReplication`** — the single-node half of the KV replication policy, and it is not the multi-node assertion with a smaller number in it. On one node one replica per bucket is the correct answer rather than the defect, so what it proves is the opposite: that nothing declared a replica count this deployment cannot place, that `spx admin kv replicas` exits 0 and reports `WANT 1`, and that `--repair` run twice against a cluster with nothing to repair changes no bucket's replica count and leaves every bucket serving. The multi-node suites cannot cover any of that, because on three nodes the unplaceable count is the right one. It lives here rather than in its own suite so it runs in every single-node nightly cell, across every host and network permutation.

 ### `natuplink` — routed NAT (`external_mode = "nat"`), one node or three

 Runs **on** a node, shelling out to `ip`/`iptables`/`ovn-nbctl` locally. Skips unless `spinifex.toml` says `external_mode = "nat"`. Nine sequential phases: host wiring (transit veth, `ip_forward`, NAT egress rules), config, the EIP surface (enabled or disabled depending on whether a public pool exists), default-subnet public IP mapping, OVN gateway and SNAT on the transit net, instance egress **proven via the serial console**, host-to-guest Tier 1 ingress, a second VPC getting a unique transit gateway IP, and the EIP ingress lifecycle.

+**The EIP ingress phase also proves hairpin traffic.** On a single node a second guest must complete a TCP handshake to the first guest's EIP on port 22, as an AWS instance reaches a public address in its own VPC. Without the `SPX-EIP-HAIRPIN` SNAT the packet returns to OVN from the transit range, which is exempt from EIP DNAT, and is never answered; found when an EKS worker on Vultr could not reach its control plane's endpoint.
+
 **Phase 9 has two shapes.** On one node it is the EIP lifecycle: allocate, associate, host delivery plumbing, a TCP handshake to the address, a vpcd restart that must replay the plumbing, and disassociate tearing it down. On a cluster it is the distributed lane instead — a guest on every node, each public IP plumbed by the node running its guest and by **no other**, every transit veth carrying the one cluster-wide MAC, each `dnat_and_snat` row carrying `external_mac` and `logical_port`, inbound proven by a handshake and outbound by each guest's own console. **That lane is the regression guard for the two defects that made routed NAT work on one node only**, and neither is visible to a single-node run.

 Phases 7 and 8 are gated on a cluster by what the datapath can actually do: the transit segment is a localnet, so host-to-guest Tier 1 ingress only exists where the VPC's gateway router port and the guest are both on this host. The VPC ingress route itself is installed on every node, so phase 8 checks it unconditionally.

-Two nightly cells run it: `nat-single` (cell 19) and `nat-multi` (cell 30). Cell 20 is the bare-metal job and is not in the permutation matrix at all, which is why the multi-node NAT cell is numbered past the end rather than filling a gap.
+Four nightly cells run it: `nat-single` (cell 19) and `nat-multi` (cell 30), and their Oracle Linux twins (cells 36 and 38). Cell 20 is the bare-metal job and is not in the permutation matrix at all, which is why the multi-node NAT cell is numbered past the end rather than filling a gap.

 ### `multinode` — behaviour that only exists on more than one node

@@ -50,6 +52,17 @@ Two nightly cells run it: `nat-single` (cell 19) and `nat-multi` (cell 30). Cell

 **`TestMultinodeDNS` convergence budgets are env-tunable.** `SPINIFEX_MULTINODE_DNS_TOPOLOGY_TIMEOUT` (default `90s`) bounds every node agreeing on the NS topology, `SPINIFEX_MULTINODE_DNS_TIMEOUT` (default `2m`) bounds a record becoming answerable from a guest or a node, and `SPINIFEX_MULTINODE_DNS_POLL` (default `3s`) is the poll interval for both. Values are Go durations; a malformed or non-positive value panics rather than falling back to the default.

+**`TestMultinodeOVNRaft` restarts each node's OVN central in turn before it kills the NB leader.**
+A `setup-ovn.sh` rerun restarts a node's NB and SB members together, on EL through the one `ovn-northd.service`, so the test does the same on every database node in turn, waits for both members to be back as cluster members under a known leader, and requires an NB write to land after each step and the full quorum with one leader at the end.
+
+**`TestMultinodeIPSec` proves guest traffic is encrypted on the wire, not only that IPsec is configured.**
+It first requires every node to carry the OVS certificate, key and CA pointers, `NB_Global.ipsec=true` never asserted over an unconfigured chassis, and an AES-GCM SA in `ip xfrm state`.
+None of those fails when SAs exist but traffic bypasses them, so it then launches two instances on distinct nodes and sends 200 pings of 1000 bytes between them.
+For each direction of that pair it fails unless the outbound SAs to the peer's `ovn-encap-ip` carried at least the payload sent, an outbound-only capture on the interface routing to the peer saw ESP (raw or UDP/4500-encapsulated), and that capture saw no plaintext Geneve (UDP/6081).
+The capture is outbound-only because inbound ESP is decrypted and seen again as Geneve on the same interface, which would read as plaintext on a healthy node.
+It proves the node pair the guests landed on; the SA check covers every node.
+It skips when the node config sets `network.ipsec_enabled=false`, or when the scheduler places all four launches on one node.
+
 ### `kvquorum` — the cluster's own state survives losing a node (multi-node, needs three)

 **The control plane keeps its state in JetStream KV buckets, and this is the suite that proves those buckets are actually replicated.** A bucket on one replica lives on one JetStream-chosen server recorded in no config, so losing that node makes it return `nats: no responders available for request` from *every* node at once — a cluster-wide outage caused by a single-node event. Leader leases are the worst case: acquiring one is a write, so a lease bucket without a quorum cannot be acquired by anyone, and every reconciler sharing it stops cluster-wide at exactly the moment a node has failed and there is repair work to do.
@@ -68,6 +81,8 @@ Three entry points, in the order they run:

 **Asserted from both sides, and the pair is the point.** `InstanceAutoRecovery` takes a node down and requires the guest to come back on exactly one survivor, identified by the qemu process rather than by what the API says, then brings the node back and requires it to drop the local copy instead of relaunching it. `InstanceRecoveryRefusesStorageFault` freezes predastore cluster-wide first, so the guest is paused by `werror=stop` when its node goes away, and requires that **nothing moves** — a store that refuses every node cannot be fixed by moving a guest, and the relaunch would cost it the request QEMU is holding. A reconciler that cannot tell a host failure from a storage failure passes the first and fails the second. Neither turns recovery on, because there is no setting for it: both wait for every node to report the reconciler active on a cluster nothing configured, so a build that reintroduced a switch fails here.

+**`InstanceRecoveryKeepsAnOlderInstanceAndItsDisk` asserts the guest survives with its data, which the first test never did.** Its guest writes a token to its root volume and prints it from the volume at every boot, and it is left running past the pending watchdog's timeout before its node is taken away, because a real instance is always older than a minute. After the survivor relaunches it, the instance must stay running for three watchdog ticks with its root volume present, and the new console must print the token, so the guest booted from the disk it wrote. The node then returns, and the instance is stopped and started once its last launch is older than the timeout again, with the same checks and a later `LaunchTime`. Before ADR-0003:S7, both relaunches carried the original launch time, and the watchdog terminated the instance seconds after QEMU started and deleted its root volume.
+
 **`InstancePartitionRecovery` is the fault this feature exists for, and the only one of the three that produces it rather than approximating it.** It drops NATS on 4222 and 4248 between one node and its peers and nothing else, so that node keeps its services, keeps its QEMU and keeps answering its own local clients while it cannot reach consensus — half healthy, which is what a real network fault looks like and what a service stop cannot imitate. It then requires that the node gives up the volume on a clock it can evaluate without reaching anything, that a survivor takes the guest over, and that the rejoining node forgets its copy rather than relaunching it. **The assertion it exists for is the one no single sample can reach:** a watch samples every node every three seconds for the whole run and fails on any sample naming two hosts, because two guests writing one volume through one object store is data loss already under way rather than a race that resolves. The evidence for *why* the guest stopped is the surrender and the fence in the node's own journal — a guest killed for any other reason satisfies the process count and means the opposite.

 **The partition gets its own nft table and a systemd deadman timer, for reasons worth keeping.** The node's real firewall is `inet spinifex_filter`, so a test that flushed or edited it would leave the node unfirewalled with no sign of which rules it lost; a separate table at priority -300 says exactly "these packets, dropped" and heals in one atomic delete. `systemd-run --on-active` removes it independently, so a cancelled workflow or a panicking test does not leave a node cut off from its cluster until somebody notices. It also proves SSH still works and proves a peer port is actually unreachable before asserting anything — a rule set that installed cleanly and dropped nothing would otherwise read as the system tolerating a partition.
@@ -104,9 +119,12 @@ CA installed and downloadable, fingerprint matches config, hostnames present in

 Internal and internet-facing ALB and NLB, HTTPS listeners, listener rules, `ModifyListener`, and NLB UDP.

+Unhealthy-target routing: an NLB with an HTTP health check on a TCP target group stops sending traffic to a target that fails it while still accepting TCP, sends traffic again once it recovers, and fails open when every target is unhealthy.
+An ALB whose targets all fail their health check keeps serving from them instead of answering 503.
+
 ### `eks`, `ecs`, `ecr` — container services

-`eks`: cluster create/delete, access entries, addon delivery, `GetToken`, kubeconfig artifacts, the EBS CSI volume path, and an IRSA pod wired with a projected SA token. `ecs`: clusters, container instances, the three network modes (`awsvpc`, `bridge`, `host`), task-role credentials fetched from the container credentials endpoint, a service behind an ELB, and ENI release on terminate. `ecr`: control-plane CRUD plus a real OCI data plane — docker login/build/push/pull round trips, multi-layer images, and raw blob HTTP.
+`eks`: cluster create/delete, access entries, addon delivery, `GetToken`, kubeconfig artifacts, the EBS CSI volume path, an IRSA web-identity exchange under the stock AWS trust policy (`StringEquals` on the issuer host-path `:sub`/`:aud`, a mismatched `:sub` denied with `AccessDenied`, a `StringLike` `:sub` granted), and an IRSA pod wired with a projected SA token. Worker IMDS is covered too: a managed nodegroup worker reports IMDSv2 required with hop limit 2, a pod with no IRSA resolves to the node role through IMDS, and a nodegroup built from a launch template takes the template's metadata options. `ecs`: clusters, container instances, the three network modes (`awsvpc`, `bridge`, `host`), task-role credentials fetched from the container credentials endpoint, a service behind an ELB, and ENI release on terminate. `ecr`: control-plane CRUD plus a real OCI data plane — docker login/build/push/pull round trips, multi-layer images, and raw blob HTTP.

 ### `rds` — managed databases

@@ -129,7 +147,6 @@ The create/describe/connect path against a live cluster: a DB VM that actually b
 - **`ddil`** — Denied, Disrupted, Intermittent, Limited network scenarios.
 - **`teardown`** — sweeps leftovers.
 - **`harness`** — shared primitives, not a suite. `harness/vpc_egress.go` carries the VPC/subnet/IGW/NAT-gateway fixtures the egress scenarios build on.
- **`manifest-check`, `manifest-lint`** — manifest validation, not cluster behaviour.

 ## Keeping this honest

@@ -0,0 +1,313 @@
+# Spinifex UI design system
+
+The console's tokens, primitives and the rules that hold pages to them.
+The live catalogue is the `/design` route: every token and primitive below rendered in both themes.
+It is served by the Vite dev server only and is absent from release builds; a release build answers it with "Page not found".
+The rules on it are proposed as `PROP-CONSOLEUI-001` in `docs/development/proposals/console-design-system/` in the mulga repository, and bind nobody until accepted.
+
+Source: `spinifex/runtime/roles/spinifexui/frontend/`.
+Tokens are in `src/styles.css`, primitives in `src/components/ui/`, and the catalogue in `src/routes/_auth/-components/design/`.
+
+## Principles
+
+- **Pages compose primitives; they do not restyle them.** If a page needs something the system lacks, add it to the system first.
+- **Colour is for status.** Surfaces are neutral, and brand orange marks the one primary action on a view.
+- **Borders, not shadows.** Overlays (dialogs, sheets, popovers, menus) are the only elevated surfaces.
+- **AWS semantics, our presentation.** States, field names and actions follow AWS.
+
+## Tokens
+
+All colours are OKLCH and come from mulgadc.com, so product and site match.
+
+### Surfaces
+
+| Token | Use |
+|---|---|
+| `background` | The app page. Warm off-white in light, ink `#14181D` in dark. |
+| `card`, `popover` | Raised surfaces. White in light, about `#1B2027` in dark. |
+| `muted`, `secondary`, `accent` | Hover rows, secondary buttons, quiet fills. |
+| `border` | Every divider and card edge. `#DCD8CE` in light, `oklch(1 0 0 / 8%)` in dark. |
+| `brand` | Brand orange `#EE4B12`, for the logo and decoration only. It fails text contrast, so it never carries text. |
+| `primary` | Darker orange in light, brighter orange in dark: primary button, link, focus ring, active tab. |
+| `sidebar` | The navigation column. |
+| `console`, `console-foreground` | A serial console log. Light grey on the light theme and near-black on the dark one, so the log follows the theme. |
+| `console-stamp`, `console-muted`, `console-ok`, `console-fail` | The log's tones: the kernel timestamp (brand yellow in dark, a darker amber in light), the subsystem or device that logged the line, and systemd's `[ OK ]` and `[FAILED]` markers. |
+
+### Status
+
+Each status has a solid token plus `-foreground` (text on a tinted background), `-subtle` (tinted background) and `-border`.
+
+| Status | Hue | AWS states |
+|---|---|---|
+| `success` | green | running, available, completed, healthy |
+| `warning` | brand yellow `#F5B500` | pending, stopping, creating, modifying, deleting, draining |
+| `danger` | cool red, hue 17 | stopped, error, failed, unhealthy, unavailable |
+| `info` | blue | informational notes |
+| `neutral` | grey `#9AA1A8` | terminated (muted text), and any unmapped state |
+
+`stopped` is danger because EC2 reads it that way, although RDS treats it as a resting state; splitting the map by service is its own change.
+
+`StateBadge` (`src/components/state-badge.tsx`) is the only mapping from an AWS state to a tone, and it exposes the result as `data-tone` for tests.
+Every status badge measures above 4.8:1 against its tinted background in both themes.
+
+**Danger is separated from the primary accent by hue alone in dark mode**: the two have the same lightness.
+So a destructive action is never a solid danger fill, and a danger state always carries a text label or an icon.
+
+### Chart palette
+
+`--chart-1` to `--chart-8` are categorical, with lightness spread so adjacent series separate under protanopia, deuteranopia and tritanopia.
+Measured with a simulated CVD ΔE, the worst pair is 7.4 in light and 7.2 in dark.
+Colour is never the only key: label every series.
+
+### Typography
+
+| Role | Font |
+|---|---|
+| Page titles and section headings (`h1`–`h3`) | Archivo (`font-heading`) |
+| UI text, tables and forms | Geist (`font-sans`) |
+| IDs, ARNs, IPs, CIDRs, code and eyebrow labels | JetBrains Mono (`font-mono`), with `tabular-nums` for figures |
+
+| Step | Size | Use |
+|---|---|---|
+| `2xl` | 26px | page title (`PageHeading`) |
+| `xl` | 22px | section heading |
+| `lg` | 17px | |
+| `base` | 15px | |
+| `sm` | 14px | body and table default |
+| `xs` | 13px | badges, secondary text |
+| `2xs` | 12px | mono eyebrow labels |
+
+The scale sits a step above Tailwind's defaults, as consoles such as Vultr's do, because 13px body text read as cramped on a wide screen.
+Code blocks are set at 12px (`--text-code-block`) so a long command still fits.
+
+### Primary and contrast
+
+The brand orange cannot be the primary: white on `#EE4B12` is 3.72:1, and as link text on a light card it is also 3.72:1, against WCAG AA's 4.5:1.
+So `primary` is its own token, and `brand` keeps the exact site colour for decoration.
+
+| Theme | `primary` | Fill text | Fill contrast | As a link |
+|---|---|---|---|---|
+| light | `oklch(55% 0.19 36.6)` | white | 5.30:1 | 5.30:1 on a card, 4.69:1 on the page |
+| dark | `oklch(68% 0.18 38)` | ink | 5.70:1 | 5.27:1 on a card |
+
+Dark mode puts ink on the fill rather than white, because no orange is both bright enough to read as a link on ink and dark enough to carry white text.
+
+### Radius
+
+`--radius` is 0.5rem; `rounded-md` is that, `rounded-sm` and `rounded-lg` step 2px either side, and status pills are `rounded-full`.
+
+### Themes
+
+The console defaults to dark, and the header switch toggles light and dark (`localStorage["spinifex-ui-theme"]`).
+`public/preboot.js` runs before the bundle. It applies the stored theme and `color-scheme` before first paint, so a dark-mode user never sees a white page while the bundle loads, and it turns off zod's eval probe (below).
+It is an external file because the console's CSP forbids inline scripts.
+
+## Constraints from the console's CSP
+
+The Go console serves `script-src 'self'; style-src 'self'`.
+
+- **No inline `<script>` or `<style>` elements.** A library that injects either breaks in production while working under `vite dev`, which applies no CSP. This is why `sonner` is not used.
+- **No runtime CSS variables set through `style`.** Primitives use fixed class maps instead, as `ToggleGroup` does for `spacing`.
+- **Base UI's `ScrollArea` injects a `<style>` element** to hide native scrollbars. `main.tsx` wraps the app in `CSPProvider disableStyleElements`, and `styles.css` carries the rule instead (`.base-ui-disable-scrollbar`).
+- **zod compiles validators with `new Function`** unless it is told not to, and it probes for that at module load, before any app code runs. `public/preboot.js` sets `globalThis.__zod_globalConfig = { jitless: true }` first.
+- **Chrome reports a violation whenever `DOMParser` meets a body that is not XML**, because the `<parsererror>` element it builds carries an inline style. The AWS SDK parses every query-protocol error body, so a proxy that answers with plain text or HTML produces a console violation as well as the error. Harmless, but it is noise in exactly the place someone debugging a failed call will look.
+- Base UI overlays animate through `data-starting-style` and `data-ending-style` transitions, not keyframe animation libraries, and every transition has `motion-reduce:transition-none`.
+
+## Primitives
+
+`src/components/ui/` holds the shadcn primitives on Base UI.
+Added in Phase 1: breadcrumb, checkbox, command (`cmdk`), dialog, hover-card, popover, resizable (`react-resizable-panels`), scroll-area, slider, switch, table, toggle and toggle-group.
+`button-group` joins buttons edge to edge, as the instance power buttons are.
+
+Every dialog closes on Escape.
+⌘↵ or Ctrl+Enter submits it: an `AlertDialog` clicks its action unless that is disabled, and a `Dialog` submits its form, including from a textarea where Enter alone adds a line.
+Enter in a typed confirmation confirms once the phrase matches.
+The confirm button shows the shortcut, ⌘↵ on a Mac and Ctrl ↵ elsewhere (`src/lib/shortcuts.ts`), and names it in `aria-keyshortcuts`.
+
+A disabled control stays visible and shows `cursor: not-allowed`, so it says why it does nothing rather than vanishing.
+An on/off setting is a `Switch`, orange when on; a `Checkbox` only picks items from a list, such as table rows, subnets or security groups.
+
+`InfoTip` (`src/components/info-tip.tsx`) is UniFi's (i): a small button beside a label that opens a short explanation on hover or click, with `InfoList` for the facts inside it.
+`PanelField` takes it through `info`, so the login user names the image's distribution and the image names its AMI details.
+
+Base UI marks orientation with `data-orientation`, so `styles.css` defines the `data-horizontal` and `data-vertical` variants the shadcn sources expect.
+
+Base UI marks the selected tab with `data-active`, not `data-selected`, so `ui/tabs` styles that attribute.
+
+shadcn's `chart` is not used: it pulls `recharts`, whose `reselect@5.1.1` dependency fails the workspace's `trustPolicy: no-downgrade`.
+Charts are uPlot instead (see Charts).
+
+## Enforcement
+
+Four lint layers keep pages on the system.
+
+| Rule | Where | What it catches |
+|---|---|---|
+| `shadcn/no-unknown-classes` | all source, not tests | a class Tailwind does not generate, such as a typo or a removed token |
+| `shadcn/no-restyle` | all source | a page passing layout or colour classes to a primitive |
+| `eslint/no-restricted-imports` | `src/routes/**` | a route importing a chart library directly |
+| `src/routes/route-files.test.ts` | test suite | a route that builds a table by hand, with `<table>` or the `ui/table` primitives |
+| `design-coverage.test.ts` | test suite | a shared component in `src/components` that `/design` neither shows nor lists with a reason |
+
+The table test allows two files, each with its reason: the catalogue's specimen of the `ui/table` primitives, and the RDS parameter editor, an editable grid.
+Every other list is a `DataTable`, and the test also fails when an allowed file stops building a table so the list cannot go stale.
+
+## Shell
+
+`src/lib/nav.ts` is the one description of the console's navigation.
+The sidebar, the command palette and the breadcrumbs all read it, and `nav.test.ts` checks every destination against the generated route tree, so a renamed route fails a test rather than leaving a dead link.
+
+- **Sidebar.** Groups take generic names rather than AWS's service marks: Compute, Databases, Authentication and Object Storage, not EC2, RDS, IAM and S3. Pages and the rest of the console keep the AWS names, and each group's `service` keeps its AWS name findable in the command palette, where typing it exactly ranks that group first. Each service group folds, and a folded group stays folded across reloads (`localStorage["spinifex-ui-nav-collapsed"]`). Navigating into a folded group unfolds it so the active item shows, and it can then be folded again, the current group included. On a phone the sidebar is a sheet, and any navigation closes it.
+- **Command palette.** ⌘K or Ctrl K, or the Search button in the header. It lists every page the user can reach, every create flow (`create` on a nav item), and a theme toggle. It is navigation only, with no resource search yet.
+- **Breadcrumbs.** Service, then list page, then the resource or the create flow. A resource ID is set in mono. On a phone only the last crumb shows.
+- **Sign in.** The sign-in card sits over the shell on a light, barely blurred overlay rather than on a page of its own, so the console stays in view. The shell behind it is `inert`, which keeps pointer, keyboard and screen reader out of it until the user signs in.
+Signing in does not reload the page: the card shrinks and fades with its veil over 300ms, then the console routes to the dashboard in place, and the page content fades in. Signing out is a 300ms view-transition cross-fade into the sign-in card, which says "Signed out." in a warning-coloured notice. Under reduced motion both switch at once.
+
+## Layout
+
+The app shell's `<main>` is `min-w-0`, so a wide table scrolls inside its card instead of widening the page.
+The page gutter is 16px on a phone, 24px from `md` and 32px from `xl`.
+`PageHeading` wraps its actions below the title when they do not fit beside it.
+Every page must fit at 390px with `document.documentElement.scrollWidth` equal to the window width.
+
+## Data tables
+
+`DataTable` (`src/components/data-table.tsx`) renders `@tanstack/react-table` over the `ui/table` primitives, so sorting, row selection and shift-click range selection are the library's and not ours to maintain.
+Pages describe columns with the small `DataColumn` shape (`id`, `header`, `cell`, `sortValue`, `hideBelow`, `align`), which the component maps to TanStack column definitions; a column with a `sortValue` sorts, with missing values last and numbers compared as numbers.
+Sort and selection stay controlled by the caller, so a page can keep them in the URL.
+
+A column hides itself through `hideBelow`, which is a **container** query on the table's card (`@max-lg`, `@max-xl`, `@max-4xl`, `@max-5xl`), not a viewport one.
+So columns drop out when a side panel narrows the list as well as on a small screen.
+
+List pages keep their filter, state and sort in the URL, so a filtered view can be bookmarked, shared and survives a reload.
+EC2 Instances is the reference page (`ec2/(instances)/-components/instances-page.tsx`):
+
+- The first column names the resource: the `Name` tag in the primary colour, with the ID in mono beneath it, opening its side panel.
+- Bulk actions appear above the table once a row is selected. Each action counts only the selected instances it applies to, so a mixed selection never sends a call that must fail, and Terminate asks first and lists every ID.
+- An empty account gets a call to action, and an empty filter result says the filters are why.
+
+VPCs and Internet gateways follow the same pattern.
+
+### Selecting rows to delete
+
+Every list with a delete action has row checkboxes. Selecting a row shows `SelectionBar` (`src/components/selection-bar.tsx`) above the table, with the count, the actions and a way to clear, and `useRowSelection` (`src/hooks/use-row-selection.ts`) keeps the selection to rows still in the list.
+`BulkDelete` (`src/components/bulk-delete.tsx`) is the bar's destructive action, and its dialog names every selected resource before anything is sent.
+
+- **One call per resource, one at a time.** Most AWS delete calls take a single ID, so each refusal lands against its own row with AWS's message. A partial failure keeps the dialog open, and its confirm becomes "Retry N" for only the ones that failed.
+- **A batch call where AWS has one.** `deleteMany` instead of `deleteOne` hands the whole selection to one function that returns the refusals by ID, as S3's `DeleteObjects` does for up to 1000 keys.
+- **No pre-filtering.** The bar sends what was selected and lets AWS refuse, rather than guessing which deletes would fail. Instances are the exception, since one call takes many IDs and a mixed selection would fail as a whole.
+- **The bulk delete and the panel's delete are the same function.** Where a single delete is several calls, such as removing dependants first, the sequence lives in one place and both use it.
+- The verb follows the AWS console: Release for an Elastic IP, Deregister for an image.
+
+### Typed confirmation
+
+A delete that cannot be walked back asks the user to type a word first, as the AWS console does before deleting a volume.
+`ConfirmPhrase` (`src/components/confirm-phrase.tsx`) is the field, and the confirm stays disabled until it matches, ignoring case and surrounding space.
+`DeleteConfirmationDialog`, `ResourceDeleteDialog` and `BulkDelete` take it through `confirmPhrase`, so the panel's delete and the bulk delete ask the same way.
+
+- **"delete"** for volumes, snapshots, RDS instances and snapshots, EKS and ECS clusters and Route 53 hosted zones. An RDS instance deleted without a final snapshot asks for its identifier instead, which is stronger.
+- **"terminate"** for instances.
+- Not for anything cheap to recreate, such as a key pair, a security group rule or a tag.
+
+Key pairs is the smallest example (`ec2/(key)/-components/key-pairs-page.tsx`).
+
+### Object browser
+
+A bucket opens as a page rather than a panel, since its contents need the width (`s3/-components/object-browser.tsx`).
+Folders and objects share one table, folders first as S3 lists them; a folder row enters the folder, and an object row opens its panel with `HeadObject`'s properties, its user metadata and a download.
+
+- **The find box is S3's prefix.** It sends `Prefix` = folder + typed text to `ListObjectsV2` after a 250ms pause, so it finds keys the current page has not loaded, and it matches from the start of the name only.
+- **Pages are S3's.** 100 entries a page with `MaxKeys`, and Next follows `NextContinuationToken`. Tokens only run forward, so the browser keeps the stack of tokens that opened each page, and Previous pops it; sorting a column sorts the page on screen.
+- **Selecting a folder selects everything under it.** Delete lists every key under the folder, then sends `DeleteObjects` in batches of 1000, and a folder that was only partly deleted says how much was left.
+- The bucket list deletes only empty buckets; S3 refuses the rest with `BucketNotEmpty`, and the dialog says so before anything is sent.
+
+#### Uploads
+
+Uploads run in a store outside React (`src/lib/uploads.ts`), so they carry on, and keep their progress, while the user browses elsewhere in the console.
+
+- **A small, quick upload stays on the button.** While a batch runs, Upload becomes "Uploading N%" with a ring that fills, and a click on it opens the progress dialog.
+- **A large or slow batch opens the dialog.** It opens at once for 16 MiB or more, and otherwise after 2 seconds if the batch is still going, so a quick upload never flashes a dialog. Hide keeps the batch on the button, and the dialog does not reopen on its own after that.
+- **The dialog shows what a user would ask about.** Bytes sent of the total, the rate in bits a second with SI prefixes (kbit/s, Mbit/s, Gbit/s) over the last 3 seconds, an approximate time left, and a row per file with its own progress and cancel. A rate that has fallen to nothing after bytes moved reads "Stalled". A finished batch gives its elapsed time and average rate, and failures can be retried.
+- **Files go three at a time, each in 8 MiB parts sent four at a time**, with lib-storage's `Upload` over `XhrHttpHandler`, since `fetch` reports no upload progress. Each file has its own client, as the handler's progress event does not say which request it came from.
+- Leaving the page while uploads run asks first, and signing out cancels them.
+- The console's S3 proxy gives each request an hour, in place of the server's 15-second read and 120-second write timeouts, which would cut off a part or a download on a slow link.
+
+### Instance launch
+
+The image picker shows each distribution's own mark from `src/assets/os/` (Ubuntu, Debian, Rocky, Alpine, CentOS), imported with `?no-inline` so it is served as a file and satisfies `img-src 'self'`. Any other distribution gets a generic server icon, and the "Other images" tile a disc.
+The instance type select states the chosen type's vCPUs and memory beneath it, from `DescribeInstanceTypes`.
+
+## Side panels
+
+Modelled on UniFi's console: a click on a row opens the resource in a panel beside the list, and a create flow opens there too.
+
+- **URL.** The panel is search state on the list route: `selected=<id>`, `tab=overview|monitoring|settings`, and `create=true`. Defaults are stripped, so a closed panel leaves the plain list address. Opening or closing a panel pushes a history entry, so Back closes it; a tab or filter change replaces the entry.
+- **Layout.** From `lg` (1024px) the panel is fixed to the right edge, full height below the header, flush with no outer margin, and it slides in over the list. The list keeps its width and columns underneath, so nothing reflows when a panel opens, and a soft gradient on the panel's left edge reads as the panel lying on top. Below `lg` there is no room to share, so the panel replaces the list as a full page with a back arrow. `useWideLayout` decides, through `matchMedia`.
+- **`DetailPanel`.** A header with the resource's own page (left) and close (right), the title and its ID, a summary of state and quick actions, then icon tabs with tooltips. Escape closes it unless a dialog or menu is open over it. Closing slides the panel out over 150ms, or at once under reduced motion. Each tab, and a create form, sits in its own error boundary, so a failed query shows in the panel and the list keeps working.
+- **Contents.** `PanelSection` is a folding card, `PanelFields` and `PanelField` a label and value list, with `copy` for a one-click copy of IDs and addresses, and `PanelTags` the tag list. Create forms take an `onClose(id?)` prop: embedded, they hide their back link and heading, and a create selects the new resource.
+- **One component is both the panel and the page.** Each resource has a single details component (`InstanceDetails`, `VpcDetails`, `InternetGatewayDetails`) that takes `layout="auto" | "page"`. The list renders it as the panel; `/ec2/describe-instances/$id` and its kind render the same component as a page, as one centred column under a standard left-aligned heading, with state and actions on the right and text tabs. A feature added to the panel is on the page by construction, and the page ignores Escape and has a back arrow in place of close. Expanding the panel to its page keeps the tab, and query results such as fetched console output come along from the cache.
+- **A click on the list's background closes the panel**, as UniFi's does. A row selects another resource instead, and links, buttons and inputs keep their own behaviour. Selecting rows for a bulk action closes it too, so the bulk bar and the panel never compete.
+- **Actions.** State actions sit in the summary as a button group, and the rest go in an Actions menu, Unikraft style. Stop, Reboot and every destructive action confirm in a dialog.
+- **Loading.** A panel or page still loading shows `PanelSpinner`, a centred spinner on the background, not a grey skeleton block.
+- **`DetailCard`, `DetailRow` and `ListCard` are retired.** Every resource page uses the panel sections instead.
+
+## Create flows
+
+Every create flow opens in the side panel over its list, through `CreatePanel`, and never in a dialog or on a page of its own.
+
+- **URL.** `create=true` opens it, plus a list-local field where the flow needs context (`volume` for a snapshot, `instance` for an RDS snapshot, `launch=service|task` on ECS clusters). The old `create-*` addresses redirect to the list with the panel open, so links and bookmarks keep working.
+- **Form.** The fields come in the order AWS asks for them. Settings that are on or off are `Switch`es, and a multi-select picker (subnets, security groups) is a `SwitchList`, a boxed list with one switch per row. Checkboxes only select table rows.
+- **CLI.** Every create form ends with a `CliCommandPanel` showing the AWS CLI commands that send the same request, built from the form's live values and only the flags the console actually sends.
+- **Submit.** `FormActions` turns the submit button into a spinner with its pending label in place, and `FormLock`, a native `<fieldset disabled>`, disables every control above it until the request settles. Nothing above the form appears or moves, and the panel does not scroll.
+- **Failure.** A refused request shows beside the button through `FormActions`'s `error`, and the form unlocks. The handler wraps only the mutation call in `try`, with an empty `catch` that returns, because react-hook-form rethrows whatever the handler throws and the error is already on screen from the mutation's state.
+- **Success.** `onClose(id)` closes the panel and selects the new resource, where the list has a panel to select it in.
+- **Appliance images.** A create that needs an ECS, EKS or RDS appliance image shows the missing-image callout inside its panel, never on the list, and only the step that needs the image is blocked.
+
+## Charts
+
+`MetricChart` (`src/components/charts/metric-chart.tsx`) draws a time series with uPlot, chosen for size and because it draws on a canvas, which the CSP allows where inline style attributes are not; `csp.e2e.ts` renders one under the production header.
+The look follows UniFi's statistics: smoothed lines over a fading fill, a light grid, a legend chip per series that reads the value under the cursor (or the latest), and `-24 hrs … Now` beneath.
+Series colours are the chart palette, and the chart redraws when the theme changes.
+
+Instance monitoring streams live statistics from the host: CPU utilisation, network bytes in and out, and packets in and out, one sample every 500 ms over the last five minutes, with no history to range over.
+The stream is `GetInstanceMetrics` on the spinifex service, a signed POST answered as `text/event-stream` (`src/lib/live-metrics.ts`), which ends every 60 s and reconnects with `Since` so nothing is lost or repeated.
+It runs only while the tab is visible and the instance is running; a segment that fails or delivers nothing backs off.
+A live chart draws stepped paths, and its x axis slides with the host's clock one second behind the newest sample (`LiveClock`), so the plot keeps moving between samples and a slow link does not jitter it.
+A gap of more than three sample periods breaks the line rather than drawing across missing time.
+
+## Dates
+
+`formatDateTime` writes the AWS console's form in the viewer's zone, `2026/06/04 02:00 GMT-5`, with a 24-hour clock and UTC as plain `GMT`.
+
+## Testing
+
+`make test-ui-e2e` rebuilds `dist` and runs `pnpm e2e`, which drives Playwright against two servers (`playwright.config.ts`):
+
+- **dev** (Vite on 3101) for behaviour: the `/design` catalogue in both themes and four widths, theme bootstrap, primitives' keyboard behaviour, the shell, and the Instances, VPC and Internet gateway list and panel flows.
+- **preview** (the built `dist` on 3102) for `csp.e2e.ts`, which serves the console's own CSP header and fails on any violation. Vite dev applies no CSP, so this is the only place an inline style or `eval` is caught before production.
+
+`e2e/console.ts` is the fixture. It seeds a fake session and answers `/proxy/**` by action, with empty defaults for what the dashboard and shell ask for; a test adds handlers for its own data, and any call nobody answered fails the test through `api.unhandled`.
+Unhandled calls get an error in the protocol's own shape, since a plain-text body would itself trip the CSP test.
+
+### Screenshots
+
+Screenshot baselines live in `e2e/__screenshots__/` and are Chromium renders taken in `en-AU` and UTC, so dates do not move with the host.
+They are drawn inside Playwright's own container image, because font rasterisation differs between hosts and a baseline drawn on a workstation fails on the runner.
+`playwright.image` in the frontend directory pins that image by version and digest, and is the one place it is named.
+Its version must equal the lockfile's `@playwright/test`, since the image carries the browsers for exactly one release, so bump both together.
+
+- `make ui-baselines` redraws every baseline in the pinned image, with `pnpm e2e --update-snapshots=all`, since the default mode keeps any baseline within the diff tolerance and so would keep one drawn on another host. Commit the changed files under `e2e/__screenshots__/` with the change that moved them.
+- `make ui-e2e-container` runs the suite in the same image and compares the screenshots, exactly as CI does.
+
+Both run `scripts/ui-playwright.sh`, which uses podman, or docker where podman is absent, and `CONTAINER_RUNTIME` overrides the choice.
+It mounts the frontend directory, installs with `pnpm install --frozen-lockfile` into a separate `node_modules` under `~/.cache/spinifex-ui-playwright`, so the host's `node_modules` is never touched, then builds `dist` and runs the suite.
+It refuses to run when the image's Node differs from `.nvmrc` or its Playwright differs from the lockfile's.
+`make test-ui-e2e` and `pnpm e2e` still run on the host, where screenshots compare reliably only against baselines drawn on that host, so do not commit baselines from `pnpm e2e:update`.
+
+The `UI End-to-End` job in `.github/workflows/ui.yml` runs the suite inside the pinned image on every pull request touching the frontend, and screenshots gate the merge with every behavioural and CSP test.
+A preceding `UI End-to-End Image` job reads `playwright.image`, so the workflow and the `make` targets cannot disagree about the image.
+On a failure the job uploads `test-results/` as the `ui-e2e-test-results` artifact, holding each failed screenshot's expected, actual and diff images and the trace.
+
+## Open
+
+- **Monitoring is live only.** Nothing is stored, so there is no `GetMetricData` and no range beyond the five-minute window; the history plan is `docs/development/feature/ec2-basic-metrics.md` in the mulga repository.
@@ -0,0 +1,417 @@
+---
+title: "Spinifex on Vultr"
+seoTitle: "Run Spinifex on Vultr with reserved IPs — Spinifex Docs"
+description: "Deploy Spinifex on Vultr VMs or bare metal with one command, and run EC2, EBS, S3, VPC, load balancers and EKS on public Vultr reserved IPs it manages."
+category: "Cloud Install"
+tags:
+  - install
+  - vultr
+  - cloud
+  - terraform
+  - cluster
+resources:
+  - title: "Setting Up Your Cluster"
+    url: "/docs/setting-up-your-cluster"
+  - title: "Spinifex on Oracle Cloud"
+    url: "/docs/oci-integration"
+  - title: "Single-Node Install"
+    url: "/docs/install"
+  - title: "Spinifex Repository"
+    url: "https://github.com/mulgadc/spinifex"
+---
+
+# Running Spinifex on Vultr
+
+> Run the AWS surface (EC2, EBS, S3, VPC, load balancers and EKS) on Vultr, with guests that get real public addresses through Vultr's own API.
+
+## Table of Contents
+
+- [Overview](#overview)
+- [Prerequisites](#prerequisites)
+- [Instructions](#instructions)
+- [1. Download Spinifex](#1-download-spinifex)
+- [2. Choose a deployment](#2-choose-a-deployment)
+- [3. Deploy](#3-deploy)
+- [4. Restrict the API key](#4-restrict-the-api-key)
+- [5. Check it worked](#5-check-it-worked)
+- [6. Obtain AWS credentials](#6-obtain-aws-credentials)
+- [7. Set up your cluster](#7-set-up-your-cluster)
+- [8. Test resiliency](#8-test-resiliency)
+- [Removing the deployment](#removing-the-deployment)
+- [Limits](#limits)
+- [Troubleshooting](#troubleshooting)
+
+---
+
+## Overview
+
+Spinifex is an open-source infrastructure platform that brings core AWS services to bare-metal, edge, and on-prem environments.
+It serves the EC2, EBS, S3, VPC, IAM and STS APIs, plus load balancers and EKS, over an endpoint of your own, so the AWS CLI, the AWS SDKs and the AWS Terraform provider all work against it unmodified.
+
+This guide runs Spinifex on Vultr, on one node or three, on VMs or on bare metal.
+It is the same product you would install on your own servers with [Single-Node Install](/docs/install), so everything you build on it behaves the same wherever Spinifex runs.
+
+**Why run it on Vultr:**
+
+- **A global network.** Vultr has 33 locations across six continents, so you can put an AWS-compatible region close to your users or your data.
+- **Compute when you need it.** VMs and bare metal are available by the hour in most Vultr locations, with no long-term commitment.
+- **Flexibility.** Start on one VM for development, then move to three nodes or bare metal for production with the same command.
+- **Portable workloads.** Your applications and Terraform code against the AWS APIs, not against Vultr, so they move between Vultr, AWS, your own rack or an edge site as a deployment decision.
+- **Keep your AWS shape.** VPCs, subnets, security groups, instances, volumes and load balancers stay exactly as you define them today.
+  No rewrite of your application or your infrastructure code is required.
+
+### How Spinifex uses your Vultr API key
+
+Spinifex includes Vultr support, which needs a Vultr API key to provide public networking.
+Whenever something you create on Spinifex needs a public address, such as an EC2 instance in a public subnet, an Elastic IP, an internet-facing load balancer or a NAT gateway, Spinifex creates a Vultr reserved IP through the API, attaches it to the node the resource runs on, and routes it to that resource.
+It deletes the reserved IP again when the resource goes away.
+
+The key is copied to the nodes over SSH and stored at `/etc/spinifex/vultr/api-key`, readable only by the Spinifex daemon.
+It is never written to the Terraform state or to the instance metadata.
+
+---
+
+## Prerequisites
+
+You need:
+
+- **A Vultr account** with room for the nodes you choose in [step 2](#2-choose-a-deployment).
+- **A Vultr API key.** In the Vultr portal, open your user menu, then **API Access**, and under your organization choose **Create API Key**.
+  Under **Access Control List**, choose **Add IP to Allowlist** and add `0.0.0.0/0` for now.
+  The nodes call the API to manage their addresses, and their addresses do not exist until the deploy creates them.
+  You replace `0.0.0.0/0` with the real addresses in [step 4](#4-restrict-the-api-key).
+- **OpenTofu or Terraform**, version 1.5 or newer. Install [OpenTofu](https://opentofu.org/docs/intro/install/) or [Terraform](https://developer.hashicorp.com/terraform/install); the deploy script uses whichever it finds, preferring OpenTofu.
+- **The AWS CLI v2** and `python3`.
+- **An SSH key pair** at `~/.ssh/id_ed25519`.
+  Create one with `ssh-keygen -t ed25519` if you do not have one.
+  The deploy installs the public key on every node.
+
+The API key page, with the key and the temporary `0.0.0.0/0` entry in its access control list:
+
+<p align="center">
+  <img src="../../.github/assets/images/vultr/vultr-api-keys.png" alt="Vultr API Access page, showing an API key and an access control list that allows 0.0.0.0/0" width="900">
+</p>
+
+---
+
+## Instructions
+
+## 1. Download Spinifex
+
+The deploy script, `deploy-vultr.sh`, is in the Spinifex repository:
+
+```bash
+git clone https://github.com/mulgadc/spinifex.git
+cd spinifex
+git switch dev
+cd scripts/terraform/vultr-spx
+```
+
+> [!NOTE]
+> Vultr support is on the development branch and not yet in a published release.
+> `git switch dev` gives you the deploy script for it, and `--channel dev` in [step 3](#3-deploy) installs the matching Spinifex build, the pre-release published from every change to `dev`.
+> Once Vultr support is in a release, skip `git switch dev` and pass `--channel latest`.
+
+---
+
+## 2. Choose a deployment
+
+| Use | Nodes | Plan | Cost while running |
+| --- | --- | --- | --- |
+| Development and testing | 1 | `vx1-g-16c-64g` VM (16 vCPU, 64 GB) | $0.48 an hour |
+| Production | 3 | `vx1-g-16c-64g` VM | $1.44 an hour |
+| Production on bare metal | 1 or 3 | `vbm-8c-64gb-amd` bare metal (8 cores, 16 threads, 64 GB, 2 × 1.9 TB NVMe) | $0.55 an hour per server |
+
+**For development and testing, use one `vx1-g-16c-64g` VM.** It is the smallest size we recommend, because the EC2 instances, load balancers, EKS and ECS clusters you create all run inside it and need headroom.
+Spinifex keeps 2 of its 16 vCPUs for the host and its own services and offers the other 14 to what you create, one vCPU each, with no overcommit.
+
+**For production, use three VMs, or bare metal.** On three nodes, an instance whose node fails is restarted on another node, and its public address follows it.
+`vbm-8c-64gb-amd` is the smallest bare-metal server we recommend for production workloads, as one server or as three.
+
+Bare metal takes longer to deliver than a VM, and Vultr bills a bare-metal server from the moment it is created, whether it is running or stopped.
+
+---
+
+## 3. Deploy
+
+`deploy-vultr.sh` is a helper script around Terraform or OpenTofu and the official [Vultr provider](https://registry.terraform.io/providers/vultr/vultr/latest), so the Vultr side of the deployment is infrastructure as code you can read, keep and change in `scripts/terraform/vultr-spx/`.
+OpenTofu or Terraform creates a private Vultr VPC and the nodes.
+The script then installs Spinifex on every node over SSH and forms the cluster.
+
+Each node runs **Ubuntu 24.04 LTS** as its host operating system, and Spinifex runs on top of it.
+Vultr installs the operating system when it creates the node, so there is no host to prepare first.
+
+Set your API key, then run the command for the deployment you chose:
+
+```bash
+export VULTR_API_KEY=<your key>
+
+# Development and testing: one VM
+./deploy-vultr.sh up --channel dev --plan vx1-g-16c-64g
+
+# Production: three VMs
+./deploy-vultr.sh up --channel dev --plan vx1-g-16c-64g --nodes 3
+
+# Production on bare metal: one server, or three
+./deploy-vultr.sh up --channel dev --plan vbm-8c-64gb-amd
+./deploy-vultr.sh up --channel dev --plan vbm-8c-64gb-amd --nodes 3
+```
+
+One VM takes about 4 minutes and three VMs about 16 minutes.
+Bare metal is slower, because Vultr takes 5 to 6 minutes to deliver a server: one server took 10 minutes in our test, and three took 18½.
+The default Vultr region is `ord` (Chicago); pass `--region` for another, and run `./deploy-vultr.sh --help` for every option.
+
+If a step fails, fix the cause and run the same command again: it resumes from the step that failed.
+
+When it finishes, the deploy prints:
+
+```ini
+endpoint=https://203.0.113.10:9999
+profile=spinifex
+ca=/home/you/.spinifex/vultr-default-ca.pem
+console=https://203.0.113.10:3000/
+node1=203.0.113.10
+api_key_allowlist=203.0.113.10/32
+```
+
+A three-node deployment lists `node2` and `node3` as well, and all three addresses in `api_key_allowlist`.
+
+---
+
+## 4. Restrict the API key
+
+Now that the node addresses are known, close the API key to everything else.
+In the Vultr portal, open your user menu, then **API Access**, and under **Access Control List**:
+
+1. Choose **Add IP to Allowlist** for each address from the `api_key_allowlist` line.
+2. Add your own machine's public address, so that `./deploy-vultr.sh` can still remove the deployment later.
+3. Remove the `0.0.0.0/0` entry, and the IPv6 `::/0` entry if your list has one.
+
+`./deploy-vultr.sh status` prints the node addresses again whenever you need them.
+
+---
+
+## 5. Check it worked
+
+```bash
+./deploy-vultr.sh status
+```
+
+This lists each node and its address, then every VM or bare-metal server, VPC, volume and reserved IP the deployment holds in your Vultr account, then the completed deploy steps.
+
+In the Vultr portal, the nodes are under **Compute → Instances**, bare-metal servers included.
+Each is named `spinifex-<deployment>-nodeN` and tagged `spinifex` and `deployment:<deployment>`, which is `deployment:default` unless you passed `--name`.
+These are the servers that run Spinifex, and the AWS workloads and services you create on it run inside them.
+
+<p align="center">
+  <img src="../../.github/assets/images/vultr/vultr-instances.png" alt="Vultr Instances page, showing Spinifex nodes named spinifex-deployment-nodeN with the spinifex and deployment tags" width="900">
+</p>
+
+To log in to a node:
+
+```bash
+./deploy-vultr.sh ssh            # node1
+./deploy-vultr.sh ssh node2      # another node of a three-node deployment
+```
+
+For a three-node deployment, check that every node has joined:
+
+```bash
+./deploy-vultr.sh ssh 'spx get nodes'
+```
+
+Expect three nodes with `Ready` status.
+
+---
+
+## 6. Obtain AWS credentials
+
+The deploy generated an AWS access key and secret on node 1.
+Display them with:
+
+```bash
+./deploy-vultr.sh credentials
+```
+
+> [!WARNING]
+> This reveals an AWS secret.
+> Do not redirect its output, or paste the credentials into chat, tickets, source control or logs.
+
+The output has this shape:
+
+```ini
+[spinifex]
+aws_access_key_id = <access-key>
+aws_secret_access_key = <secret-key>
+```
+
+Use these to sign in to the web console at the `console` address the deploy printed, `https://<node1 address>:3000/`.
+
+The deploy has also written them to an AWS profile named `spinifex` on your machine, with the endpoint and the cluster's certificate already set, so the AWS CLI works straight away:
+
+```bash
+export AWS_PROFILE=spinifex
+aws ec2 describe-instance-types
+```
+
+The same profile is set up on node 1, if you prefer to work from there:
+
+```bash
+./deploy-vultr.sh ssh 'export AWS_PROFILE=spinifex; aws ec2 describe-instance-types'
+```
+
+If this returns a list of available instance types, your installation is working.
+
+<p align="center">
+  <img src="../../.github/assets/spinifex-ui.jpg" alt="Spinifex Web UI" width="900">
+</p>
+
+**Congratulations, Spinifex is installed.**
+
+---
+
+## 7. Set up your cluster
+
+Spinifex is running, but it holds nothing yet: no machine images, no networks, no instances.
+
+Continue to [Setting Up Your Cluster](/docs/setting-up-your-cluster) to import an AMI, create an SSH key pair, create a VPC with a public subnet, and launch your first instance.
+An instance you launch into a public subnet gets a Vultr reserved IP, which you can see in the Vultr portal under **Network → Reserved IPs**.
+
+---
+
+## 8. Test resiliency
+
+A three-node deployment keeps running when a node fails.
+Spinifex restarts the failed node's instances on the surviving nodes, and moves each instance's Vultr reserved IP with it, so its public address keeps working.
+To watch it happen, run a load-balanced website across the three nodes, then power one of them off.
+
+### Launch a website on every node
+
+Import the Ubuntu image the web servers run on:
+
+```bash
+./deploy-vultr.sh ssh 'spx admin images import --name ubuntu-26.04-x86_64'
+```
+
+The [Nginx Web Server (Load Balanced)](/docs/nginx-alb) workbook launches web servers behind an Application Load Balancer.
+With these two variables it launches three, and a spread placement group puts each one on a different node:
+
+```bash
+cd docs/terraform-workbooks/nginx-alb
+export AWS_PROFILE=spinifex
+tofu init
+tofu apply -var region=us-east-1 \
+  -var spinifex_endpoint="$(aws configure get endpoint_url)" \
+  -var server_count=3 -var placement_strategy=spread
+```
+
+Use `terraform` in place of `tofu` if that is what you installed.
+Check that each web server is on its own node:
+
+```bash
+../../../scripts/terraform/vultr-spx/deploy-vultr.sh ssh 'spx get vms'
+```
+
+The `NODE` column shows `node1`, `node2` and `node3` once each, and the load balancer is the `sys.micro` instance.
+
+### Keep requests running
+
+Get the load balancer's address, and request the page once a second in a second terminal:
+
+```bash
+ALB_IP=$(aws elbv2 describe-load-balancers --names nginx-alb \
+  --query 'LoadBalancers[0].AvailabilityZones[].LoadBalancerAddresses[].IpAddress' --output text)
+while :; do
+  printf '%s %s\n' "$(date +%T)" "$(curl -s -m2 "http://$ALB_IP/" | grep -o 'Server [0-9]' || echo FAILED)"
+  sleep 1
+done
+```
+
+Each line names the web server that answered.
+
+### Power off a node
+
+From the `vultr-spx` directory, power off a node that runs a web server but not the load balancer:
+
+```bash
+./deploy-vultr.sh stop node3
+```
+
+This is a hard power-off through the Vultr API, as a hardware failure would be, so the node hands nothing over first.
+Watch the requests and `spx get vms`:
+
+1. For up to about a minute, a few requests fail while the load balancer's health checks find that one web server is gone.
+2. After that, the other two web servers answer every request.
+3. About three minutes after the power-off, the lost web server is `running` on one of the surviving nodes.
+4. About four minutes after the power-off, its own public address answers again, and the load balancer sends it requests once its health checks pass.
+
+Power the node back on:
+
+```bash
+./deploy-vultr.sh start node3
+```
+
+It rejoined the cluster about a minute after a VM started, and about five minutes after a bare-metal server started.
+Wait until `spx get nodes` shows it `Ready` before you power off another node.
+The instances that moved stay where they are.
+
+### What we measured
+
+On three `vx1-g-16c-64g` VMs and on three `vbm-8c-64gb-amd` bare-metal servers in Chicago, with the workbook above:
+
+| Node powered off | Deployment | Instance `running` elsewhere | Its public address answers | Load balancer requests failed |
+| --- | --- | --- | --- | --- |
+| A node with one web server | VMs | 173 s | 244 s | 5 of 91, all within 45 s |
+| A node with one web server | Bare metal | 172 s | 223 s | 6 of 83, all within 60 s |
+| node1, with a web server, the load balancer and the cluster's NATS leader | VMs | 167 s | 253 s | 22 of 66, all within 115 s |
+| A node with two web servers and the load balancer | Bare metal | 204 s | 271 s | 21 of 79, all within 114 s |
+
+Each case was measured once.
+The requests column counts one-second probes, so a run with more timed-out requests sends fewer of them.
+The load balancer recovers first, because it moves to a surviving node like any other instance.
+While node1 is off, the AWS endpoint in your `spinifex` profile does not answer, because it is node1's address.
+Call another node meanwhile by adding `--endpoint-url https://<node2-address>:9999` to an `aws` command.
+
+When you are done, remove the workbook with `tofu destroy` and the same `-var` options before [removing the deployment](#removing-the-deployment).
+
+---
+
+## Removing the deployment
+
+Remove anything you built on the cluster first, then:
+
+```bash
+./deploy-vultr.sh destroy
+```
+
+It shows what it will delete and asks you to type the deployment name, `default`, to confirm.
+It deletes the nodes and the VPC, plus anything else the deployment left in your Vultr account that keeps billing: VM boot volumes, and any reserved IP Spinifex created.
+When it finishes, nothing of the deployment is left, and if anything is, it lists it and fails.
+
+After removing the deployment, remove the node addresses from the API key's **Access Control** list, since Vultr may give them to someone else.
+
+---
+
+## Limits
+
+- **One or three nodes.** Other cluster sizes are not supported.
+- **16 vCPU per node is the recommended minimum.** Spinifex keeps 1 CPU on a node with up to 8, 2 on one with up to 16 and 4 on one with up to 63, and offers the rest to instances one to one, so a 16-vCPU node offers 14 and an 8-vCPU node 7.
+- **A deleted reserved IP keeps its quota slot for 24 hours**, by Vultr's documentation, so creating and deleting many addresses in a day can reach the account's limit.
+- **A three-node deployment survives one failed node at a time.**
+  With two of the three nodes down, the cluster loses quorum, and in our test the instances on the remaining node went to `error` and stayed there after a second node returned.
+- **A recovered instance can share a node with another member of its spread placement group**, when only one node has room for it.
+  It stays there after the failed node returns.
+- **Disassociating an Elastic IP leaves the instance with no public address**, where AWS would give it a new one.
+  Stopping and starting the instance gets it one.
+
+---
+
+## Troubleshooting
+
+| Symptom | Cause and fix |
+| --- | --- |
+| `the API key was refused; check its IP allowlist` | Your machine's address is not in the key's **Access Control** list. Add it in the portal. |
+| `the node's API calls are refused` or `never logged its Vultr allocator` | A node's address is not in the key's **Access Control** list. Add `0.0.0.0/0` until the deploy finishes, then follow [step 4](#4-restrict-the-api-key), and run `up` again. |
+| `plan ... is not available in this region` | Pick another region with `--region`, or check the plan's availability in the Vultr portal. |
+| `formation failed` | The full log is in `~/.spinifex/vultr/default/form.log`. Run `up` again once the cause is fixed. |
+| Any other failure | The logs are under `~/.spinifex/vultr/default/`, and the message names the log on the node. `up` resumes from the failed step. |
@@ -4,12 +4,15 @@ go 1.27.3

 require (
 	github.com/aws/aws-sdk-go v1.55.8
-	github.com/aws/aws-sdk-go-v2 v1.47.3
-	github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.23
+	github.com/aws/aws-sdk-go-v2 v1.47.4
+	github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.24
 	github.com/aws/aws-sdk-go-v2/config v1.33.9
 	github.com/aws/aws-sdk-go-v2/credentials v1.20.9
 	github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.20.3
+	github.com/aws/aws-sdk-go-v2/service/lambda v1.113.3
 	github.com/aws/aws-sdk-go-v2/service/s3 v1.114.4
+	github.com/aws/aws-sdk-go-v2/service/sqs v1.52.6
+	github.com/aws/smithy-go v1.28.6
 	github.com/cenkalti/backoff/v4 v4.3.0
 	github.com/charmbracelet/bubbles v1.0.0
 	github.com/charmbracelet/bubbletea v1.3.10
@@ -30,6 +33,7 @@ require (
 	github.com/mulgadc/viperblock v1.23.0
 	github.com/nats-io/nats-server/v2 v2.15.1
 	github.com/nats-io/nats.go v1.54.0
+	github.com/nats-io/nkeys v0.4.16
 	github.com/opencontainers/runtime-spec v1.3.0
 	github.com/oracle/oci-go-sdk/v65 v65.128.0
 	github.com/ovn-kubernetes/libovsdb v0.9.0
@@ -64,8 +68,8 @@ require (
 	github.com/Microsoft/hcsshim v0.15.0-rc.4 // indirect
 	github.com/antithesishq/antithesis-sdk-go v0.8.0-default-no-op // indirect
 	github.com/atotto/clipboard v0.1.4 // indirect
-	github.com/aws/aws-sdk-go-v2/internal/configsources v1.5.6 // indirect
-	github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.8.6 // indirect
+	github.com/aws/aws-sdk-go-v2/internal/configsources v1.5.7 // indirect
+	github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.8.7 // indirect
 	github.com/aws/aws-sdk-go-v2/internal/v4a v1.5.6 // indirect
 	github.com/aws/aws-sdk-go-v2/service/internal/accept-encoding v1.13.21 // indirect
 	github.com/aws/aws-sdk-go-v2/service/internal/checksum v1.11.7 // indirect
@@ -75,7 +79,6 @@ require (
 	github.com/aws/aws-sdk-go-v2/service/sso v1.38.4 // indirect
 	github.com/aws/aws-sdk-go-v2/service/ssooidc v1.43.4 // indirect
 	github.com/aws/aws-sdk-go-v2/service/sts v1.51.4 // indirect
-	github.com/aws/smithy-go v1.28.6 // indirect
 	github.com/aymanbagabas/go-osc52/v2 v2.0.1 // indirect
 	github.com/beorn7/perks v1.0.1 // indirect
 	github.com/boltdb/bolt v1.3.1 // indirect
@@ -159,7 +162,6 @@ require (
 	github.com/muesli/termenv v0.16.0 // indirect
 	github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 // indirect
 	github.com/nats-io/jwt/v2 v2.8.2 // indirect
-	github.com/nats-io/nkeys v0.4.16 // indirect
 	github.com/nats-io/nuid v1.0.1 // indirect
 	github.com/opencontainers/go-digest v1.0.0 // indirect
 	github.com/opencontainers/image-spec v1.1.1 // indirect
@@ -22,20 +22,20 @@ github.com/atotto/clipboard v0.1.4 h1:EH0zSVneZPSuFR11BlR9YppQTVDbh5+16AmcJi4g1z
 github.com/atotto/clipboard v0.1.4/go.mod h1:ZY9tmq7sm5xIbd9bOK4onWV4S6X0u6GY7Vn0Yu86PYI=
 github.com/aws/aws-sdk-go v1.55.8 h1:JRmEUbU52aJQZ2AjX4q4Wu7t4uZjOu71uyNmaWlUkJQ=
 github.com/aws/aws-sdk-go v1.55.8/go.mod h1:ZkViS9AqA6otK+JBBNH2++sx1sgxrPKcSzPPvQkUtXk=
-github.com/aws/aws-sdk-go-v2 v1.47.3 h1:HRoT/MkNi3bjO5xF4dCzLzvVmkvRGHj3/cLINDnJejk=
-github.com/aws/aws-sdk-go-v2 v1.47.3/go.mod h1:xSl+OSI5ZUw6bSMupWc218zE5Pl+65MCCu3ySWFY99A=
-github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.23 h1:/fGPpKrXTjzkhsDXExBgNT2cGwxiF9rLrQXh8RlXMhI=
-github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.23/go.mod h1:QWUDuRi9PFjBHIU2iH1Pkldbe7zqLfXJC4wWLbF8nuQ=
+github.com/aws/aws-sdk-go-v2 v1.47.4 h1:bTvk6UdSSSvaQsfLVj1JRDEEqj/PDRwGf1KanaRzau0=
+github.com/aws/aws-sdk-go-v2 v1.47.4/go.mod h1:XkrjAZvr705Nytv6PkPUhoSB2W9+vVHBp5oAWpkh8Ak=
+github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.24 h1:t7FxqJ5KHDeKceyD4HtGaG6z7hMSeVGjs5/4c+L5ArM=
+github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.24/go.mod h1:5XK6NRVlK8s8yj4zKcM43HLluGA/rAz+azLLvWDMRr0=
 github.com/aws/aws-sdk-go-v2/config v1.33.9 h1:Hv6bcQUJVnmuWaKjqD6Ot8sE75NrWsWV20oGT5aWle0=
 github.com/aws/aws-sdk-go-v2/config v1.33.9/go.mod h1:ivowmrZNfBQLcg4QkaItgi5wae4Bzot9RzIdbY1hfTg=
 github.com/aws/aws-sdk-go-v2/credentials v1.20.9 h1:SNO55PnWhaDafsOcv16lCA0kdoFMl/YlegWoVsU02s0=
 github.com/aws/aws-sdk-go-v2/credentials v1.20.9/go.mod h1:+2reOTRtdD3VSBqpRDNwrZIYy8yMCYa7oK3WKTJ/I1I=
 github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.20.3 h1:gr3b87OLB29jymWPzp/CAyXwQRJVE+sC4Wm0qBIKHsk=
 github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.20.3/go.mod h1:BRjf4L/Yx2hY54o+ZibPbLJyMMeSwqYQXwcJfqW66a8=
-github.com/aws/aws-sdk-go-v2/internal/configsources v1.5.6 h1:CqL8NN3p4I0JTwwV5MEoatIl4Y2amZ2X9dRD+0hbRL0=
-github.com/aws/aws-sdk-go-v2/internal/configsources v1.5.6/go.mod h1:exdJ32hMqXl0Wyo6NwqEC5yIhtua3+PDaKE8P29m4qE=
-github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.8.6 h1:FpBq24TG6WVSObTNhQOv7Q0088rounUUlwn3zlNItkM=
-github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.8.6/go.mod h1:ONzdPVYC95oBOu0VWNEsKEntcoemP3xVKjT+qWml1s4=
+github.com/aws/aws-sdk-go-v2/internal/configsources v1.5.7 h1:fSNAt6hs4qLjb9tMZc6vj3PHlOCYGXK2JjJZ7AW2tf4=
+github.com/aws/aws-sdk-go-v2/internal/configsources v1.5.7/go.mod h1:TMav+EuXQu5kNpZZz/6rIC8xx4mCsW+6AKHrjvOSAsE=
+github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.8.7 h1:7NS6kqBjkWIN8P/AEdmtyq6eQGAhN1oHTZdzXclpH6Y=
+github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.8.7/go.mod h1:1PFqtAUsIgF1nEVojzRJSf/v88zjzqrGW+w+UAtkXyQ=
 github.com/aws/aws-sdk-go-v2/internal/v4a v1.5.6 h1:Qmn60RWuk9+bd2ddvJJqwf3Cpac3VvGMOdbGemtUhX4=
 github.com/aws/aws-sdk-go-v2/internal/v4a v1.5.6/go.mod h1:h/YJhe4AVyQxEUMluph10iqV7mjajH7A9L+BHF5ppis=
 github.com/aws/aws-sdk-go-v2/service/internal/accept-encoding v1.13.21 h1:e6YW5uME30TkrElR+4/xcY7NB6CCj2dcv7vtmGXhfhQ=
@@ -46,10 +46,14 @@ github.com/aws/aws-sdk-go-v2/service/internal/presigned-url v1.14.6 h1:LHHdobjo5
 github.com/aws/aws-sdk-go-v2/service/internal/presigned-url v1.14.6/go.mod h1:CBGB0evAiaMD3Lk4wPO7LQCcCZLuhzL5Qf/wfEUHQ5w=
 github.com/aws/aws-sdk-go-v2/service/internal/s3shared v1.20.6 h1:mYViDr2azcgxOpRrObZOl631bvavqILoE/L+Kxgbwys=
 github.com/aws/aws-sdk-go-v2/service/internal/s3shared v1.20.6/go.mod h1:OXVdAo9jNhggB+hJRGOyy7WolGTKB6Rg7IIulDeSzPM=
+github.com/aws/aws-sdk-go-v2/service/lambda v1.113.3 h1:oSHldse7GDHpdq81XeAJnhIbsdDKP9bMIt00G8W6uXI=
+github.com/aws/aws-sdk-go-v2/service/lambda v1.113.3/go.mod h1:Y//K046w4uhiR3o77beo/y52Jv2uEY13iJ+3+jicLO8=
 github.com/aws/aws-sdk-go-v2/service/s3 v1.114.4 h1:MrkWRB58Qw5TEkb51QVooVLjlM3IevO2iVXEz74B8pg=
 github.com/aws/aws-sdk-go-v2/service/s3 v1.114.4/go.mod h1:yfIxaKRHasvfy6tNEm3/LpqKsupMzQjMaWj9Y0jMVbo=
 github.com/aws/aws-sdk-go-v2/service/signin v1.10.4 h1:tvpDjuUvgKURf3Yk4hOaGbZ4QdY+070OWcA5d6LUv2A=
 github.com/aws/aws-sdk-go-v2/service/signin v1.10.4/go.mod h1:3ibVKg5Ng82Zby8++jELTD2nHxiPk2DJRLEeW/LKp8M=
+github.com/aws/aws-sdk-go-v2/service/sqs v1.52.6 h1:Suq1LNhMBpsmShEv3vO+J3H8HEeK+GwYMJDYLE4IJSs=
+github.com/aws/aws-sdk-go-v2/service/sqs v1.52.6/go.mod h1:rHGuzFg4AfVnWZa3/533tn5GQd1LzVkTXI9WMT8LqAc=
 github.com/aws/aws-sdk-go-v2/service/sso v1.38.4 h1:ChCtagixSFOxltv0o6E3NT8F7ksvZtZGPwUBynO5Nmk=
 github.com/aws/aws-sdk-go-v2/service/sso v1.38.4/go.mod h1:+FEO9jrVMpc9oN3CdAxJbtlQEu3z8UddDPSedpeM3PY=
 github.com/aws/aws-sdk-go-v2/service/ssooidc v1.43.4 h1:j6c+HA4AoS0EBEjArPjavFEBrucJEryfrQU0TffmACk=
@@ -77,8 +77,8 @@ func TestEC2ErrorCatalogMetadataAndClasses(t *testing.T) {
 	if catalog.metadata.VerifiedOn == "" {
 		t.Fatal("verifiedOn is empty")
 	}
-	if len(catalog.codes) != 114 {
-		t.Fatalf("catalog code count = %d, want 114", len(catalog.codes))
+	if len(catalog.codes) != 115 {
+		t.Fatalf("catalog code count = %d, want 115", len(catalog.codes))
 	}
 	if catalog.codes["InvalidParameterValue"] != ec2ErrorClassClient {
 		t.Fatal("InvalidParameterValue is not classified as client")
@@ -17,6 +17,7 @@ var modelFiles = map[Service]string{
 	ElasticLoadBalancingV2: "models/elastic-load-balancing-v2-2015-12-01.json.gz",
 	IAM:                    "models/iam-2010-05-08.json.gz",
 	RDS:                    "models/rds-2014-10-31.json.gz",
+	Route53:                "models/route-53-2013-04-01.json.gz",
 	S3:                     "models/s3-2006-03-01.json.gz",
 	STS:                    "models/sts-2011-06-15.json.gz",
 }
@@ -1,8 +1,10 @@
 package awsmodel

 import (
+	"bytes"
 	"encoding/base64"
 	"encoding/json"
+	"encoding/xml"
 	"fmt"
 	"maps"
 	"net/http"
@@ -39,7 +41,7 @@ func (r *EncodedRequest) URL(endpoint string) string {
 }

 // EncodeRequest serializes input, as produced by GenerateRequests, for the
-// service's protocol. S3's rest-xml is not supported.
+// service's protocol: query, ec2, json, rest-json or rest-xml.
 func EncodeRequest(service Service, operationName string, input map[string]any) (*EncodedRequest, error) {
 	model, err := Load(service)
 	if err != nil {
@@ -57,6 +59,8 @@ func EncodeRequest(service Service, operationName string, input map[string]any)
 		return encoder.json(operation, input)
 	case "rest-json":
 		return encoder.restJSON(operation, input)
+	case "rest-xml":
+		return encoder.restXML(operation, input)
 	default:
 		return nil, fmt.Errorf("awsmodel: encoding %s requests is not implemented", model.metadata.Protocol)
 	}
@@ -282,18 +286,50 @@ func (e requestEncoder) jsonValue(ref ShapeRef, value any) (any, error) {
 }

 func (e requestEncoder) restJSON(operation *Operation, input map[string]any) (*EncodedRequest, error) {
+	request, members, err := e.restBindings(operation, input)
+	if err != nil || len(members) == 0 {
+		return request, err
+	}
+	shape := e.model.shapes[operation.Input.Shape]
+	var document any
+	if payload, ok := members[shape.Payload]; ok && shape.Payload != "" {
+		if document, err = e.jsonValue(shape.Members[shape.Payload], payload); err != nil {
+			return nil, err
+		}
+	} else {
+		body := make(map[string]any, len(members))
+		for member, value := range members {
+			ref := shape.Members[member]
+			encoded, err := e.jsonValue(ref, value)
+			if err != nil {
+				return nil, err
+			}
+			body[elementName(member, ref)] = encoded
+		}
+		document = body
+	}
+	if request.Body, err = json.Marshal(document); err != nil {
+		return nil, err
+	}
+	request.Header.Set("Content-Type", "application/json")
+	return request, nil
+}
+
+// restBindings places the URI, query string and header members of input on a
+// new request, and returns the members left for the body.
+func (e requestEncoder) restBindings(operation *Operation, input map[string]any) (*EncodedRequest, map[string]any, error) {
 	request := &EncodedRequest{Method: operation.HTTP.Method, Query: url.Values{}, Header: http.Header{}}
 	path := operation.HTTP.RequestURI
 	if operation.Input == nil {
 		request.Path = path
-		return request, nil
+		return request, nil, nil
 	}
 	shape := e.model.shapes[operation.Input.Shape]
 	body := map[string]any{}
 	for _, member := range slices.Sorted(maps.Keys(input)) {
 		ref, ok := shape.Members[member]
 		if !ok {
-			return nil, fmt.Errorf("%s has no member %q", operation.Input.Shape, member)
+			return nil, nil, fmt.Errorf("%s has no member %q", operation.Input.Shape, member)
 		}
 		value := input[member]
 		target := e.model.shapes[ref.Shape]
@@ -301,18 +337,18 @@ func (e requestEncoder) restJSON(operation *Operation, input map[string]any) (*E
 		case "uri":
 			text, err := scalarString(value, ref, target, "iso8601")
 			if err != nil {
-				return nil, fmt.Errorf("%s: %w", member, err)
+				return nil, nil, fmt.Errorf("%s: %w", member, err)
 			}
 			path = strings.Replace(path, "{"+member+"+}", escapeGreedyLabel(text), 1)
 			path = strings.Replace(path, "{"+member+"}", url.PathEscape(text), 1)
 		case "querystring":
 			if err := e.restQuery(request.Query, ref, target, value); err != nil {
-				return nil, fmt.Errorf("%s: %w", member, err)
+				return nil, nil, fmt.Errorf("%s: %w", member, err)
 			}
 		case "header":
 			text, err := e.headerValue(ref, target, value)
 			if err != nil {
-				return nil, fmt.Errorf("%s: %w", member, err)
+				return nil, nil, fmt.Errorf("%s: %w", member, err)
 			}
 			request.Header.Set(ref.LocationName, text)
 		case "headers":
@@ -321,41 +357,222 @@ func (e requestEncoder) restJSON(operation *Operation, input map[string]any) (*E
 				request.Header.Set(ref.LocationName+key, fmt.Sprint(entry))
 			}
 		default:
-			encoded, err := e.jsonValue(ref, value)
-			if err != nil {
-				return nil, err
-			}
-			name := member
-			if ref.LocationName != "" {
-				name = ref.LocationName
-			}
-			body[name] = encoded
+			body[member] = value
 		}
 	}
 	request.Path = path
-	if payload := shape.Payload; payload != "" {
-		if document, ok := body[payload]; ok {
-			body = nil
-			encoded, err := json.Marshal(document)
-			if err != nil {
-				return nil, err
-			}
-			request.Body = encoded
-		}
+	return request, body, nil
+}
+
+// restXML encodes the body as AWS SDKs do: a root element named for the input
+// shape, or for the payload member, in the service's namespace. An input with
+// body members sends its root element even when none of them is set.
+func (e requestEncoder) restXML(operation *Operation, input map[string]any) (*EncodedRequest, error) {
+	request, members, err := e.restBindings(operation, input)
+	if err != nil || operation.Input == nil {
+		return request, err
 	}
-	if len(body) > 0 {
-		encoded, err := json.Marshal(body)
-		if err != nil {
+	shape := e.model.shapes[operation.Input.Shape]
+	var root *xmlElement
+	switch {
+	case shape.Payload != "":
+		value, ok := members[shape.Payload]
+		if !ok {
+			return request, nil
+		}
+		ref := shape.Members[shape.Payload]
+		target := e.model.shapes[ref.Shape]
+		if target.Type != "structure" {
+			switch value := value.(type) {
+			case []byte:
+				request.Body = value
+			case string:
+				request.Body = []byte(value)
+			default:
+				return nil, fmt.Errorf("%s: want a blob or string payload, got %T", shape.Payload, value)
+			}
+			return request, nil
+		}
+		name := ref.LocationName
+		if name == "" {
+			name = elementName(ref.Shape, ShapeRef{LocationName: target.LocationName})
+		}
+		if root, err = e.xmlRoot(name, ref, value); err != nil {
 			return nil, err
 		}
-		request.Body = encoded
+	case hasBodyMembers(shape):
+		name := elementName(operation.Input.Shape, ShapeRef{LocationName: shape.LocationName})
+		if root, err = e.xmlRoot(name, *operation.Input, members); err != nil {
+			return nil, err
+		}
+	default:
+		return request, nil
 	}
-	if request.Body != nil {
-		request.Header.Set("Content-Type", "application/json")
+	var body bytes.Buffer
+	encoder := xml.NewEncoder(&body)
+	if err := root.encode(encoder); err != nil {
+		return nil, err
 	}
+	if err := encoder.Flush(); err != nil {
+		return nil, err
+	}
+	request.Body = body.Bytes()
+	request.Header.Set("Content-Type", "application/xml")
 	return request, nil
 }

+func hasBodyMembers(shape *Shape) bool {
+	for _, ref := range shape.Members {
+		if ref.Location == "" {
+			return true
+		}
+	}
+	return false
+}
+
+// xmlRoot is the structure value encoded under name, in its shape's namespace
+// or else the service's.
+func (e requestEncoder) xmlRoot(name string, ref ShapeRef, value any) (*xmlElement, error) {
+	holder := &xmlElement{}
+	if err := e.xmlValue(holder, name, ref, value); err != nil {
+		return nil, err
+	}
+	root := holder.children[0]
+	if namespace := e.model.metadata.XMLNamespace; namespace.URI != "" && e.model.shapes[ref.Shape].XMLNamespace.URI == "" {
+		root.attrs = append([]xml.Attr{namespaceAttr(namespace)}, root.attrs...)
+	}
+	return root, nil
+}
+
+func (e requestEncoder) xmlStructure(parent *xmlElement, shapeName string, fields map[string]any) error {
+	shape := e.model.shapes[shapeName]
+	for _, member := range slices.Sorted(maps.Keys(fields)) {
+		ref, ok := shape.Members[member]
+		if !ok {
+			return fmt.Errorf("%s has no member %q", shapeName, member)
+		}
+		name := elementName(member, ref)
+		if ref.XMLAttribute {
+			text, err := scalarString(fields[member], ref, e.model.shapes[ref.Shape], "iso8601")
+			if err != nil {
+				return fmt.Errorf("%s: %w", name, err)
+			}
+			parent.attrs = append(parent.attrs, xml.Attr{Name: xml.Name{Local: name}, Value: text})
+			continue
+		}
+		if err := e.xmlValue(parent, name, ref, fields[member]); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+// xmlValue appends the elements value is encoded as to parent. A list or map
+// is wrapped in an element named name unless it is flattened.
+func (e requestEncoder) xmlValue(parent *xmlElement, name string, ref ShapeRef, value any) error {
+	shape := e.model.shapes[ref.Shape]
+	switch shape.Type {
+	case "structure":
+		fields, ok := value.(map[string]any)
+		if !ok {
+			return fmt.Errorf("%s: want a structure, got %T", name, value)
+		}
+		element := parent.add(name)
+		if shape.XMLNamespace.URI != "" {
+			element.attrs = append(element.attrs, namespaceAttr(shape.XMLNamespace))
+		}
+		return e.xmlStructure(element, ref.Shape, fields)
+	case "list":
+		items, ok := value.([]any)
+		if !ok {
+			return fmt.Errorf("%s: want a list, got %T", name, value)
+		}
+		container, itemName := parent, name
+		if !ref.Flattened {
+			container, itemName = parent.add(name), elementName("member", *shape.Member)
+		}
+		for _, item := range items {
+			if err := e.xmlValue(container, itemName, *shape.Member, item); err != nil {
+				return err
+			}
+		}
+		return nil
+	case "map":
+		entries, ok := value.(map[string]any)
+		if !ok {
+			return fmt.Errorf("%s: want a map, got %T", name, value)
+		}
+		container, entryName := parent, name
+		if !ref.Flattened {
+			container, entryName = parent.add(name), "entry"
+		}
+		keyName, valueName := elementName("key", *shape.Key), elementName("value", *shape.Value)
+		for _, key := range slices.Sorted(maps.Keys(entries)) {
+			entry := container.add(entryName)
+			entry.add(keyName).text = key
+			if err := e.xmlValue(entry, valueName, *shape.Value, entries[key]); err != nil {
+				return err
+			}
+		}
+		return nil
+	default:
+		text, err := scalarString(value, ref, shape, "iso8601")
+		if err != nil {
+			return fmt.Errorf("%s: %w", name, err)
+		}
+		parent.add(name).text = text
+		return nil
+	}
+}
+
+// elementName is a member's wire name: its xmlName or jsonName, else its own.
+func elementName(member string, ref ShapeRef) string {
+	if ref.LocationName != "" {
+		return ref.LocationName
+	}
+	return member
+}
+
+func namespaceAttr(namespace XMLNamespace) xml.Attr {
+	name := "xmlns"
+	if namespace.Prefix != "" {
+		name += ":" + namespace.Prefix
+	}
+	return xml.Attr{Name: xml.Name{Local: name}, Value: namespace.URI}
+}
+
+// xmlElement is one element of an encoded rest-xml body.
+type xmlElement struct {
+	name     string
+	attrs    []xml.Attr
+	text     string
+	children []*xmlElement
+}
+
+func (x *xmlElement) add(name string) *xmlElement {
+	child := &xmlElement{name: name}
+	x.children = append(x.children, child)
+	return child
+}
+
+func (x *xmlElement) encode(encoder *xml.Encoder) error {
+	start := xml.StartElement{Name: xml.Name{Local: x.name}, Attr: x.attrs}
+	if err := encoder.EncodeToken(start); err != nil {
+		return err
+	}
+	if x.text != "" {
+		if err := encoder.EncodeToken(xml.CharData(x.text)); err != nil {
+			return err
+		}
+	}
+	for _, child := range x.children {
+		if err := child.encode(encoder); err != nil {
+			return err
+		}
+	}
+	return encoder.EncodeToken(start.End())
+}
+
 func (e requestEncoder) restQuery(query url.Values, ref ShapeRef, shape *Shape, value any) error {
 	switch shape.Type {
 	case "list":
@@ -0,0 +1,212 @@
+package awsmodel_test
+
+import (
+	"cmp"
+	"encoding/xml"
+	"io"
+	"net/url"
+	"slices"
+	"strings"
+	"testing"
+
+	"github.com/aws/aws-sdk-go/aws"
+	"github.com/aws/aws-sdk-go/aws/credentials"
+	"github.com/aws/aws-sdk-go/aws/request"
+	"github.com/aws/aws-sdk-go/aws/session"
+	"github.com/aws/aws-sdk-go/service/route53"
+	"github.com/aws/aws-sdk-go/service/s3"
+	"github.com/mulgadc/spinifex/internal/awsmodel"
+	"github.com/stretchr/testify/require"
+)
+
+const sdkEndpoint = "https://aws.example"
+
+var sdkSession = session.Must(session.NewSession(&aws.Config{
+	Region:           aws.String("us-east-1"),
+	Endpoint:         aws.String(sdkEndpoint),
+	Credentials:      credentials.AnonymousCredentials,
+	S3ForcePathStyle: aws.Bool(true),
+}))
+
+// TestEncodeRequestRestXMLMatchesSDK checks the encoder against aws-sdk-go's
+// restxml marshaller for the same input, so the expected wire form is the
+// SDK's rather than one written by hand.
+func TestEncodeRequestRestXMLMatchesSDK(t *testing.T) {
+	r53 := route53.New(sdkSession)
+	tests := []struct {
+		name      string
+		service   awsmodel.Service
+		operation string
+		input     map[string]any
+		sdk       *request.Request
+	}{
+		{
+			name:      "CreateHostedZone",
+			service:   awsmodel.Route53,
+			operation: "CreateHostedZone",
+			input: map[string]any{
+				"Name":             "example.com",
+				"CallerReference":  "ref-1",
+				"HostedZoneConfig": map[string]any{"Comment": "a & b", "PrivateZone": true},
+				"VPC":              map[string]any{"VPCRegion": "us-east-1", "VPCId": "vpc-1"},
+			},
+			sdk: requestOf(r53.CreateHostedZoneRequest(&route53.CreateHostedZoneInput{
+				Name:             aws.String("example.com"),
+				CallerReference:  aws.String("ref-1"),
+				HostedZoneConfig: &route53.HostedZoneConfig{Comment: aws.String("a & b"), PrivateZone: aws.Bool(true)},
+				VPC:              &route53.VPC{VPCRegion: aws.String("us-east-1"), VPCId: aws.String("vpc-1")},
+			})),
+		},
+		{
+			name:      "ChangeResourceRecordSets",
+			service:   awsmodel.Route53,
+			operation: "ChangeResourceRecordSets",
+			input: map[string]any{
+				"HostedZoneId": "Z0123456789",
+				"ChangeBatch": map[string]any{
+					"Comment": "two changes",
+					"Changes": []any{
+						map[string]any{"Action": "UPSERT", "ResourceRecordSet": map[string]any{
+							"Name": "www.example.com", "Type": "A", "TTL": int64(300),
+							"ResourceRecords": []any{map[string]any{"Value": "192.0.2.1"}, map[string]any{"Value": "192.0.2.2"}},
+						}},
+						map[string]any{"Action": "DELETE", "ResourceRecordSet": map[string]any{
+							"Name": "old.example.com", "Type": "TXT", "TTL": int64(60),
+							"ResourceRecords": []any{map[string]any{"Value": `"v=1"`}},
+						}},
+					},
+				},
+			},
+			sdk: requestOf(r53.ChangeResourceRecordSetsRequest(&route53.ChangeResourceRecordSetsInput{
+				HostedZoneId: aws.String("Z0123456789"),
+				ChangeBatch: &route53.ChangeBatch{
+					Comment: aws.String("two changes"),
+					Changes: []*route53.Change{
+						{Action: aws.String("UPSERT"), ResourceRecordSet: &route53.ResourceRecordSet{
+							Name: aws.String("www.example.com"), Type: aws.String("A"), TTL: aws.Int64(300),
+							ResourceRecords: []*route53.ResourceRecord{{Value: aws.String("192.0.2.1")}, {Value: aws.String("192.0.2.2")}},
+						}},
+						{Action: aws.String("DELETE"), ResourceRecordSet: &route53.ResourceRecordSet{
+							Name: aws.String("old.example.com"), Type: aws.String("TXT"), TTL: aws.Int64(60),
+							ResourceRecords: []*route53.ResourceRecord{{Value: aws.String(`"v=1"`)}},
+						}},
+					},
+				},
+			})),
+		},
+		{
+			name:      "ChangeTagsForResource",
+			service:   awsmodel.Route53,
+			operation: "ChangeTagsForResource",
+			input: map[string]any{
+				"ResourceType":  "hostedzone",
+				"ResourceId":    "Z0123456789",
+				"AddTags":       []any{map[string]any{"Key": "team", "Value": "dns"}, map[string]any{"Key": "env", "Value": "prod"}},
+				"RemoveTagKeys": []any{"old", "stale"},
+			},
+			sdk: requestOf(r53.ChangeTagsForResourceRequest(&route53.ChangeTagsForResourceInput{
+				ResourceType:  aws.String("hostedzone"),
+				ResourceId:    aws.String("Z0123456789"),
+				AddTags:       []*route53.Tag{{Key: aws.String("team"), Value: aws.String("dns")}, {Key: aws.String("env"), Value: aws.String("prod")}},
+				RemoveTagKeys: []*string{aws.String("old"), aws.String("stale")},
+			})),
+		},
+		{
+			name:      "UpdateHostedZoneComment without a comment",
+			service:   awsmodel.Route53,
+			operation: "UpdateHostedZoneComment",
+			input:     map[string]any{"Id": "Z0123456789"},
+			sdk:       requestOf(r53.UpdateHostedZoneCommentRequest(&route53.UpdateHostedZoneCommentInput{Id: aws.String("Z0123456789")})),
+		},
+		{
+			name:      "ListResourceRecordSets",
+			service:   awsmodel.Route53,
+			operation: "ListResourceRecordSets",
+			input:     map[string]any{"HostedZoneId": "Z0123456789", "StartRecordName": "www.example.com", "StartRecordType": "A", "MaxItems": int64(5)},
+			sdk: requestOf(r53.ListResourceRecordSetsRequest(&route53.ListResourceRecordSetsInput{
+				HostedZoneId: aws.String("Z0123456789"), StartRecordName: aws.String("www.example.com"), StartRecordType: aws.String("A"), MaxItems: aws.String("5"),
+			})),
+		},
+		{
+			name:      "DeleteHostedZone",
+			service:   awsmodel.Route53,
+			operation: "DeleteHostedZone",
+			input:     map[string]any{"Id": "Z0123456789"},
+			sdk:       requestOf(r53.DeleteHostedZoneRequest(&route53.DeleteHostedZoneInput{Id: aws.String("Z0123456789")})),
+		},
+		{
+			name:      "S3 structure payload",
+			service:   awsmodel.S3,
+			operation: "PutBucketTagging",
+			input: map[string]any{
+				"Bucket":  "bucket",
+				"Tagging": map[string]any{"TagSet": []any{map[string]any{"Key": "team", "Value": "storage"}}},
+			},
+			sdk: requestOf(s3.New(sdkSession).PutBucketTaggingRequest(&s3.PutBucketTaggingInput{
+				Bucket:  aws.String("bucket"),
+				Tagging: &s3.Tagging{TagSet: []*s3.Tag{{Key: aws.String("team"), Value: aws.String("storage")}}},
+			})),
+		},
+	}
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			require.NoError(t, test.sdk.Build())
+			want := test.sdk.HTTPRequest
+			wantBody, err := io.ReadAll(test.sdk.GetBody())
+			require.NoError(t, err)
+
+			got, err := awsmodel.EncodeRequest(test.service, test.operation, test.input)
+			require.NoError(t, err)
+			require.Equal(t, want.Method, got.Method)
+			gotURL, err := url.Parse(got.URL(sdkEndpoint))
+			require.NoError(t, err)
+			// The SDK's api-2.json ends ChangeResourceRecordSets' path in a
+			// slash that the Smithy model drops; the gateway accepts both.
+			require.Equal(t, strings.TrimSuffix(want.URL.Path, "/"), gotURL.Path)
+			require.Equal(t, want.URL.Query(), gotURL.Query())
+			if len(wantBody) == 0 {
+				require.Empty(t, got.Body)
+				require.Empty(t, got.Header.Get("Content-Type"))
+				return
+			}
+			require.Equal(t, canonicalXML(t, wantBody), canonicalXML(t, got.Body))
+			require.Equal(t, "application/xml", got.Header.Get("Content-Type"))
+		})
+	}
+}
+
+func requestOf[O any](r *request.Request, _ O) *request.Request { return r }
+
+// xmlTree is a parsed XML element. The SDK emits sibling elements in map
+// order, so canonicalXML sorts them by name, keeping list items in order.
+type xmlTree struct {
+	XMLName  xml.Name   `xml:""`
+	Attrs    []xml.Attr `xml:",any,attr"`
+	Text     string     `xml:",chardata"`
+	Children []xmlTree  `xml:",any"`
+}
+
+func canonicalXML(t *testing.T, body []byte) xmlTree {
+	t.Helper()
+	var tree xmlTree
+	require.NoError(t, xml.Unmarshal(body, &tree), string(body))
+	var canonicalise func(*xmlTree)
+	canonicalise = func(node *xmlTree) {
+		node.Text = strings.TrimSpace(node.Text)
+		slices.SortFunc(node.Attrs, func(a, b xml.Attr) int { return cmp.Compare(a.Name.Space+a.Name.Local, b.Name.Space+b.Name.Local) })
+		slices.SortStableFunc(node.Children, func(a, b xmlTree) int { return cmp.Compare(a.XMLName.Local, b.XMLName.Local) })
+		for i := range node.Children {
+			canonicalise(&node.Children[i])
+		}
+	}
+	canonicalise(&tree)
+	return tree
+}
+
+func TestEncodeRequestRestXMLPutsRootInServiceNamespace(t *testing.T) {
+	request, err := awsmodel.EncodeRequest(awsmodel.Route53, "CreateHostedZone", map[string]any{"Name": "example.com", "CallerReference": "ref"})
+	require.NoError(t, err)
+	tree := canonicalXML(t, request.Body)
+	require.Equal(t, xml.Name{Space: "https://route53.amazonaws.com/doc/2013-04-01/", Local: "CreateHostedZoneRequest"}, tree.XMLName)
+	require.Equal(t, "/2013-04-01/hostedzone", request.Path)
+}
@@ -117,8 +117,3 @@ func TestEncodeRequestRejectsUnmodelledMember(t *testing.T) {
 	_, err := awsmodel.EncodeRequest(awsmodel.IAM, "GetRole", map[string]any{"RoleName": "app", "Bogus": "x"})
 	require.ErrorContains(t, err, `no member "Bogus"`)
 }
-
-func TestEncodeRequestS3IsNotSupported(t *testing.T) {
-	_, err := awsmodel.EncodeRequest(awsmodel.S3, "ListBuckets", map[string]any{})
-	require.ErrorContains(t, err, "not implemented")
-}
@@ -119,7 +119,7 @@ func (m *Model) decodeErrorCode(body []byte) (string, error) {
 			return "", fmt.Errorf("JSON error envelope has no code")
 		}
 		return code, nil
-	case "ec2", "query":
+	case "ec2", "query", "rest-xml":
 		root, err := parseXML(body)
 		if err != nil {
 			return "", err
@@ -123,3 +123,16 @@ func TestValidateErrorResponseClassifiesCommonErrorsAsUnmodelled(t *testing.T) {
 		})
 	}
 }
+
+func TestDecodeRequestResultReadsRestXMLErrorEnvelope(t *testing.T) {
+	body := []byte(`<?xml version="1.0"?>
+<ErrorResponse xmlns="https://route53.amazonaws.com/doc/2013-04-01/"><Error><Type>Sender</Type><Code>NoSuchHostedZone</Code><Message>No hosted zone found with ID: Z1</Message></Error><RequestId>r-1</RequestId></ErrorResponse>`)
+	result := DecodeRequestResult(Route53, 404, nil, body)
+	if result.Code != "NoSuchHostedZone" || result.Message != "No hosted zone found with ID: Z1" {
+		t.Fatalf("result = %+v", result)
+	}
+	violations, modelled, err := ValidateErrorResponse(Route53, "GetHostedZone", body)
+	if err != nil || !modelled || len(violations) != 0 {
+		t.Fatalf("ValidateErrorResponse = %v, %v, %v; want a declared code", violations, modelled, err)
+	}
+}
@@ -34,6 +34,7 @@ const (
 	ElasticLoadBalancingV2 Service = "elasticloadbalancingv2"
 	IAM                    Service = "iam"
 	RDS                    Service = "rds"
+	Route53                Service = "route53"
 	S3                     Service = "s3"
 	STS                    Service = "sts"
 )
@@ -50,6 +51,14 @@ type Metadata struct {
 	ServiceID       string
 	SigningName     string
 	TargetPrefix    string
+	XMLNamespace    XMLNamespace
+}
+
+// XMLNamespace is the namespace an XML element is serialized in. URI is empty
+// when none is declared.
+type XMLNamespace struct {
+	URI    string `json:"uri"`
+	Prefix string `json:"prefix"`
 }

 // Operation describes an AWS API operation and the shapes used by its input,
@@ -84,6 +93,7 @@ type Shape struct {
 	Pattern         string
 	LocationName    string
 	TimestampFormat string
+	XMLNamespace    XMLNamespace
 	Payload         string
 	Flattened       bool
 	Sensitive       bool
@@ -5,6 +5,6 @@ package awsmodel
 // ModelCommit is the github.com/aws/api-models-aws commit the embedded
 // models were taken from, and ModelCommitDate the day it was committed.
 const (
-	ModelCommit     = "9b1a4929274443d8671ea8acfda7afbe856e29e9"
-	ModelCommitDate = "2026-09-25"
+	ModelCommit     = "e10d6a1c5ac36af7c0c61d1f67b3c3ebbe41de49"
+	ModelCommitDate = "2026-10-28"
 )
@@ -8,7 +8,7 @@ import (
 )

 func TestServices(t *testing.T) {
-	want := []Service{ACM, EC2, ECR, ECS, EKS, ElasticLoadBalancingV2, IAM, RDS, S3, STS}
+	want := []Service{ACM, EC2, ECR, ECS, EKS, ElasticLoadBalancingV2, IAM, RDS, Route53, S3, STS}
 	if got := Services(); !reflect.DeepEqual(got, want) {
 		t.Fatalf("Services() = %v, want %v", got, want)
 	}
@@ -21,13 +21,14 @@ func TestLoadAllModels(t *testing.T) {
 		shapes     int
 	}{
 		ACM:                    {"2015-12-08", 40, 239},
-		EC2:                    {"2016-11-15", 804, 4245},
-		ECR:                    {"2015-09-21", 58, 404},
-		ECS:                    {"2014-11-13", 77, 643},
-		EKS:                    {"2017-11-01", 70, 450},
+		EC2:                    {"2016-11-15", 811, 4297},
+		ECR:                    {"2015-09-21", 62, 426},
+		ECS:                    {"2014-11-13", 88, 734},
+		EKS:                    {"2017-11-01", 70, 458},
 		ElasticLoadBalancingV2: {"2015-12-01", 51, 418},
 		IAM:                    {"2010-05-08", 180, 624},
-		RDS:                    {"2014-10-31", 164, 791},
+		RDS:                    {"2014-10-31", 168, 816},
+		Route53:                {"2013-04-01", 71, 409},
 		S3:                     {"2006-03-01", 112, 710},
 		STS:                    {"2011-06-15", 11, 101},
 	}
--- a/Show More
+++ b/Show More