acme-server: cert-manager integration test + production hardening (Phase 5/7)

Closes the production-readiness loop on the ACME surface. After this
commit, certctl ships per-account rate limits + a GC sweeper for
expired ACME state + a kind-driven cert-manager 1.15 integration test
+ a lego-driven RFC conformance harness + a k6 loadtest scenario for
the unauthenticated ACME path.

Architecture:
  - Rate limits live in-memory + per-replica. Restart wipes the
    counters; orders/hour caps are eventual-consistency anyway. A
    3-replica certctl-server fleet behind an LB effectively has 3x
    the configured throughput per account; persistent rate limiting
    is a follow-up if production telemetry shows abuse patterns we
    can't catch in a single restart cycle. Per-key + per-action
    isolation: ActionNewOrder/acc-1, ActionKeyChange/acc-1, and
    ActionChallengeRespond/<challenge-id> are independent buckets.
  - GC loop follows the existing scheduler-loop pattern (atomic.Bool
    + sync.WaitGroup; see crlGenerationLoop for shape). Three
    independent SQL sweeps per tick (DELETE expired nonces; UPDATE
    pending authzs whose expires_at < now() to expired; UPDATE
    pending/ready/processing orders whose expires_at < now() to
    invalid). Each sweep is a single statement; failures are logged-
    and-continued so a failing nonces sweep doesn't block authzs.
    Per-sweep 1m timeout bounds a stuck Postgres.
  - cert-manager integration test is gated on KIND_AVAILABLE so CI
    skips it cleanly (kind is too heavy for per-PR). Operators run
    locally via 'make acme-cert-manager-test'; the harness brings up
    a fresh cluster each run + tears it down on Cleanup.
  - lego conformance harness drives a real ACME client through
    register → run → cert-PEM-landed against a hermetic certctl
    stack. Catches RFC-shape regressions third-party clients would
    hit before they ship.
  - k6 ACME-flow scenario hammers the unauthenticated surface
    (directory + new-nonce + ARI synthetic-id) at 100 VUs × 5m. JWS-
    signed flows are out of scope for k6 (no JWS support); they're
    covered by the lego harness above.

What ships:
  - internal/api/acme/ratelimit.go (+ ratelimit_test.go: 7 cases —
    disable-when-perHour-zero, capacity, per-key isolation, per-
    action isolation, refill-over-time, RetryAfter, concurrent-access
    with -race + 200 goroutines × 200 calls).
  - internal/repository/postgres/acme.go: 4 new methods —
    CountActiveOrdersByAccount + GCExpiredNonces + GCExpireAuthorizations
    + GCInvalidateExpiredOrders. Each a single SQL statement.
  - internal/service/acme.go: SetRateLimiter + GarbageCollect +
    rate-limit gates at 3 entry points (CreateOrder + RotateAccountKey
    + RespondToChallenge) + concurrent-orders gate at CreateOrder.
    2 new sentinels (ErrACMERateLimited, ErrACMEConcurrentOrdersExceeded);
    5 new GC metrics (gc_runs / gc_run_failures / gc_nonces_reaped /
    gc_authzs_expired / gc_orders_invalidated).
  - internal/scheduler/scheduler.go: ACMEGarbageCollector interface +
    acmeGCRunning atomic.Bool + acmeGCInterval + 2 setters (SetACME-
    GarbageCollector + SetACMEGCInterval) + acmeGCLoop following the
    crlGenerationLoop shape.
  - internal/api/handler/acme.go: writeServiceError gains rateLimited
    (429 + RFC 8555 §6.7) + concurrent-orders-exceeded mappings.
  - internal/config/config.go: 5 new env vars
    (CERTCTL_ACME_SERVER_RATE_LIMIT_ORDERS_PER_HOUR=100,
    CERTCTL_ACME_SERVER_RATE_LIMIT_CONCURRENT_ORDERS=5,
    CERTCTL_ACME_SERVER_RATE_LIMIT_KEY_CHANGE_PER_HOUR=5,
    CERTCTL_ACME_SERVER_RATE_LIMIT_CHALLENGE_RESPONDS_PER_HOUR=60,
    CERTCTL_ACME_SERVER_GC_INTERVAL=1m).
  - cmd/server/main.go: NewRateLimiter() + SetRateLimiter() at
    startup; conditional SetACMEGarbageCollector(acmeService) +
    SetACMEGCInterval(cfg.ACMEServer.GCInterval) when Enabled+
    GCInterval > 0.
  - deploy/test/acme-integration/: kind-config.yaml + cert-manager-
    install.sh + clusterissuer-trust-authenticated.yaml +
    clusterissuer-challenge.yaml + certificate-test.yaml + conformance-
    lego.sh + certmanager_test.go (//go:build integration + KIND_AVAILABLE
    gate).
  - deploy/test/loadtest/k6/acme_flow.js + README ACME-flows section.
  - Makefile: 2 new PHONY targets (acme-cert-manager-test +
    acme-rfc-conformance-test).
  - docs/acme-server.md: status flipped to Phase 5; Configuration
    table grows 5 rows; new 'Phase 5 — operational guidance' section
    explaining rate-limit math + GC sweeper semantics + cert-manager
    integration + lego conformance + k6 baseline.

Tests:
  - 'go vet ./...' clean across the repo.
  - 'go test -short -count=1 ./internal/...' green across every
    affected package (service / acme / handler / scheduler / repo /
    config).
  - 'go vet -tags=integration ./deploy/test/acme-integration/' clean
    (the integration test compiles cleanly with the build tag).
  - The kind/cert-manager harness is gated behind KIND_AVAILABLE so
    CI skips by default; operators run locally via 'make acme-cert-
    manager-test'.

Engineering history: cowork/WORKSPACE-CHANGELOG.md 'ACME-Server-5'.

deploy/test/loadtest/README.md

@@ -313,7 +313,47 @@ deploy/test/loadtest/
 └── results/          (gitignored — k6 writes summary.{json,txt} here)
 ```
 
+## ACME flows (Phase 5)
+
+The `deploy/test/loadtest/k6/acme_flow.js` scenario hammers the
+unauthenticated ACME surface (directory + new-nonce + ARI synthetic
+lookups) at constant 100 VUs for 5 minutes. JWS-signed paths
+(new-account / new-order / finalize) are intentionally out of scope:
+k6 doesn't ship JWS, and bundling lego inside k6 would obscure the
+underlying-server p95 we're trying to measure. Instead, the
+`make acme-rfc-conformance-test` target drives lego against the same
+stack for the full happy-path conformance gate.
+
+Run it:
+
+```
+cd deploy/test/loadtest
+docker compose up -d certctl postgres
+k6 run --env CERTCTL_ACME_DIRECTORY=https://localhost:8443/acme/profile/prof-test/directory \
+       k6/acme_flow.js
+```
+
+### Baseline (ACME flows, 100 VUs × 5m)
+
+The baseline is operator-captured on a workstation-class machine with
+a single certctl-server container + a single postgres container.
+Re-capture after schema migrations or transport changes; commit the
+new numbers so regressions are visible in code review.
+
+| Metric                                     | Threshold | Last captured | Notes |
+|--------------------------------------------|-----------|---------------|-------|
+| `directory_duration` p95                   | < 500 ms  | _operator_    | Unauth GET; cache-friendly. |
+| `new_nonce_duration` p95                   | < 300 ms  | _operator_    | Single Postgres INSERT under the hood. |
+| `renewal_info_duration` p95 (synthetic id) | < 800 ms  | _operator_    | Synthetic cert-id → 4xx fast path. |
+| `http_req_failed` rate                     | < 1%      | _operator_    | Should be ~0 — failures here mean transport issues. |
+
+Capture command: `make loadtest` after pointing the compose stack at
+the ACME flow scenario. Operators with kind / cert-manager available
+should pair this with `make acme-cert-manager-test` for end-to-end
+verification.
+
 ## Audit references
 
 - API tier:       `cowork/issuer-coverage-audit-2026-05-01/RESULTS.md` fix #8.
 - Connector tier: `cowork/deployment-target-audit-2026-05-02/RESULTS.md` Bundle 10.
+- ACME flows:     Phase 5 master prompt (`cowork/acme-server-prompts/06-phase-5-certmanager-hardening-prompt.md`).

deploy/test/loadtest/k6/acme_flow.js

@@ -0,0 +1,80 @@
+// Phase 5 — k6 scenario for the ACME issuance loop. Each VU executes
+// directory + new-nonce + new-account + new-order + finalize + cert
+// download against an operator-provided certctl-server. Per-step
+// duration histograms feed the baseline numbers in
+// deploy/test/loadtest/README.md (ACME flows section).
+//
+// Default scenario: 100 concurrent VUs for 5 minutes. Override via
+// K6_VUS / K6_DURATION env vars.
+//
+// Note on signing: this scenario runs as a *load* generator, not as a
+// JWS-signing client. It exercises the unauthenticated surface
+// (directory + new-nonce + GET renewal-info) and validates that the
+// server holds throughput under concurrency. JWS-signed flow load is
+// a follow-up that requires bundling lego or a dedicated Go driver
+// inside the k6 binary — k6 itself doesn't ship JWS.
+
+import http from "k6/http";
+import { check, sleep } from "k6";
+import { Trend } from "k6/metrics";
+
+const directoryURL =
+  __ENV.CERTCTL_ACME_DIRECTORY ||
+  "https://certctl:8443/acme/profile/prof-test/directory";
+
+export const options = {
+  scenarios: {
+    acme_directory_and_nonce: {
+      executor: "constant-vus",
+      vus: parseInt(__ENV.K6_VUS || "100", 10),
+      duration: __ENV.K6_DURATION || "5m",
+      gracefulStop: "30s",
+    },
+  },
+  insecureSkipTLSVerify: true, // self-signed bootstrap cert
+  thresholds: {
+    "directory_duration": ["p(95)<500"],
+    "new_nonce_duration": ["p(95)<300"],
+    "renewal_info_duration": ["p(95)<800"],
+    "http_req_failed": ["rate<0.01"],
+  },
+};
+
+const directoryDuration = new Trend("directory_duration", true);
+const newNonceDuration = new Trend("new_nonce_duration", true);
+const renewalInfoDuration = new Trend("renewal_info_duration", true);
+
+export default function () {
+  // Step 1 — directory.
+  let res = http.get(directoryURL);
+  directoryDuration.add(res.timings.duration);
+  check(res, { "directory 200": (r) => r.status === 200 });
+
+  if (res.status !== 200) return;
+  const dir = res.json();
+
+  // Step 2 — new-nonce.
+  if (dir.newNonce) {
+    res = http.head(dir.newNonce);
+    newNonceDuration.add(res.timings.duration);
+    check(res, {
+      "new-nonce 200 + Replay-Nonce": (r) =>
+        r.status === 200 && !!r.headers["Replay-Nonce"],
+    });
+  }
+
+  // Step 3 — ARI smoke (with a deliberately-malformed cert-id to
+  // exercise the error path; full happy-path needs a real cert which
+  // requires JWS signing — out of scope for this baseline scenario).
+  if (dir.renewalInfo) {
+    res = http.get(dir.renewalInfo + "/" + "aaaa.bbbb");
+    renewalInfoDuration.add(res.timings.duration);
+    // 400 (malformed cert-id, expected) OR 404 (cert not found).
+    check(res, {
+      "renewal-info 4xx for synthetic cert-id": (r) =>
+        r.status === 400 || r.status === 404,
+    });
+  }
+
+  sleep(1);
+}