mirror of
https://github.com/rcourtman/Pulse.git
synced 2026-10-23 20:25:56 +00:00
217ebc3db5b609c5f5914e264450856dff7813b2
6208 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 rcourtman
|
217ebc3db5 |
Remove the unreachable Workloads chart visibility toggle (#2734)
The Workloads View disclosure rendered `ChartVisibilityToggleButton`
only inside `<Show when={props.onChartsToggle}>`, and no surface passes
`onChartsToggle` or `chartsCollapsed` since the workload summary chart
section was deleted with the summary components. The button had no other
production consumer.
This removes:
- the `Summary` block and import in `WorkloadsFilter.tsx`, and the
`chartsCollapsed` / `onChartsToggle` props in `workloadsFilterModel.ts`
- `ChartVisibilityToggleButton` and its icon import in
`shared/FilterToolbar.tsx`
- the `chart-visibility-toggle-button` rule and
`chart-visibility-local-toggle-labels` pattern guard in
`scripts/shared-template-registry.json`
- the unit tests for the button and the WorkloadsFilter charts describe
The guardrail case that required the primitive now keeps the exact
retired registry ids, export and props out. It deliberately does not
block unrelated future chart controls: a returning summary section is a
product decision that should register its own primitive.
Contract prose: `frontend-primitives.md` (View-disclosure list, the
FilterToolbar rule, the registry paragraph, the FilterBar view-options
list), `performance-and-scalability.md` (the workload chart-visibility
sentence) and `unified-resources.md` rule 12 now say no chart show/hide
control exists instead of describing the deleted one.
Verification:
- frontend `src/features`, `src/components/shared`,
`src/components/Workloads`, `src/utils` suites on pulse-dev: 783 files
passed, one load-induced 5s timeout in `AgentsMachinesTable.test.tsx`
that passes alone (32/32) and does not reference these symbols; tsc,
prettier and the shared-template audit pass
- browser: production builds of this change and of main's three runtime
files, served against the 50-node mock rig, Proxmox and vSphere
overviews at 1440x900 and 390x844 (phone via Filters > View). The View
preferences disclosure shows identical headings, controls and pressed
states before and after, with no Summary heading, no Charts button, no
horizontal overflow and no console errors. The after bundle no longer
carries the `Show charts` copy.
|
||
|
rcourtman
|
b9c98a4a16 |
Stop Patrol fetching the circuit-breaker status nobody reads (#2704)
The Patrol page fetched `GET /api/ai/circuit/status` on every load, every 60 s poll, every return to the tab, and again in the background after Retry. The store's breaker state had one production reader: the `circuitBreakerStatus` memo in `usePatrolIntelligenceState`, which nothing read and which #2695 removes. An open breaker already reaches the page through `/api/ai/patrol/status` as the `circuit_open` blocked cause, and the paused banner shows its reason. ## Change - `stores/aiIntelligence.ts`: remove the breaker signal, getter, `circuitBreakerStatusSignal`, `loadCircuitBreakerStatus()`, its org-switch reset, and its `loadDashboardData()` entry. - `features/patrol/usePatrolIntelligenceState.ts`: drop the read from the Retry path's background supporting reads. - `api/ai.ts`: remove `AIAPI.getCircuitBreakerStatus` and the `CircuitBreakerStatus` type; nothing else used them. The backend route stays. - Tests: the client's unit tests and the store and page test mocks go. The store bundle test still pins the exact AIAPI read set. The breaker route stubs are removed from seven E2E specs. - Contracts: patrol-intelligence, ai-runtime, api-contracts and frontend-primitives say the browser reads no breaker status and how an open breaker still reaches Patrol. Each note sits next to #2677's correlation note. ## Evidence - Normalized production-build compare: only the removed store loader, signal, getter and client differ in `index.js`, plus the removed supporting read in `AIIntelligence.js`. `check-bundle-size.mjs` passes. - Private demo-mode stacks for the parent and this tree, driven by the same Playwright script at 1440x900 and 390x844. Across Patrol load, tab hide/show and stale-data Retry, the breaker read drops from 1-2 requests per phase to none. Every other endpoint matches, the Open in Assistant handoff is unchanged, and neither build logs a console error. - A forced `circuit_open` Patrol status still shows "Patrol paused / circuit breaker is open" on both builds. - `tsc --noEmit`, eslint, and vitest for `src/stores`, `src/api`, `src/features/patrol`, `src/components/AI`, the Patrol page test and the boundaries test pass, as does the full pre-commit battery. Stacked on #2695 (which stacks on #2687) with #2677 merged in, so until those land this PR shows their commits too. The 2677 owner agreed this read belongs in a follow-up. `git merge-tree` is clean against main, #2677, #2690, #2691 and #2693. |
||
|
rcourtman
|
533027ee06 |
Drop Patrol state members no surface reads (#2695)
`usePatrolIntelligenceState()` returned 18 members that no file outside the hook reads. A TypeScript checker scan of src and tests found no readers and no spread, computed or `Object.*` escapes of the state object. Several were also dead inside the hook. ## Change - **Return object.** Drops activePatrolFindings, activityRefreshTrigger, acceptedManualRunId, assistantHandoffFindingId, circuitBreakerStatus, fullModeUnlocked, initialSurfaceReady, isRefreshing, loadAllData, openAssistantOperationsLoopForFinding, the four decision/outcome counts, both work-evidence counts, recordPatrolWorkflowStarterActivity and setFullModeUnlocked, plus handleAssistantFindingHandoff (see below). - **Dead locals.** Removes the write-only signals (initialSurfaceReady, isRefreshing, fullModeUnlocked, assistantHandoffFindingId), the 15 s background refresh timeout that only reset isRefreshing, the unread memos, and the never-called `openAssistantOperationsLoopForFinding`. That last one was a second copy of the handoff `FindingsPanel` already performs. Locals still used inside the hook (activePatrolFindings, loadAllData, activityRefreshTrigger, acceptedManualRunId) stay. - **Test-only exports.** `openPatrolAssistantWorkflowHandoff` and `recordPatrolWorkflowStarterActivity` were used only by that dead path and by their own tests, so they go. - **Workspace.** `handleAssistantFindingHandoff` only wrote the unread handoff signal, so `PatrolIntelligenceWorkspace` no longer passes `FindingsPanel` an `onAssistantHandoff` callback. The prop is optional. - **Tests.** Raw-source pins on removed code are dropped. The manual-sync timeout test now pins the Retry path's own generation guard; a mutation of its timeout fails it. The non-blocking starter test now exercises `recordPatrolControlStarterActivity`. - **Contracts.** - patrol-intelligence said a finding-to-Assistant handoff *must* record a `pulse_patrol` workflow starter. Only the dead path did that; the live `FindingsPanel` handoff never has, so `pulse_patrol` has no browser producer. The clause now says that and keeps the content-free rules as a condition. - The "local Patrol state must expose work evidence" sentence no longer requires counts nothing reads. - Three refresh clauses that described a header refresh action now name the stale-data Retry, the only operator refresh left. - api-contracts and frontend-primitives record that `FindingsPanel` owns the handoff. ## Not in this PR The store still fetches `GET /api/ai/circuit/status` for nothing. That removal is a follow-up after #2677, because it edits the same lines. ## Evidence - Production builds before and after compared with chunk hashes, SRI strings and short identifiers normalised. Only the removed hook code differs, the unused `pulse_patrol` constant drops out of `aiChat.js`, and `index.js` reorders the same preload list. `check-bundle-size.mjs` passes. - Private demo-mode stacks for both builds, driven by the same Playwright script at 1440x900 and 390x844: Patrol load, Activity > Finding options and history, Open in Assistant (drawer opens with the finding), tab hide/show, and the forced stale-data banner plus Retry. Each phase hit the same `/api` endpoints on both builds, and neither build had console errors. - `tsc --noEmit`, eslint, and vitest for `src/features/patrol`, `AIIntelligence.test.tsx`, `frontendResourceTypeBoundaries.test.ts` and `FindingsPanel.test.ts` pass, as does the full pre-commit battery. Stacked on #2687 (which carries #2682 and #2686). Until those merge, this PR shows their commits too. `git merge-tree` is clean against main, #2690, #2677 and #2691. |
||
|
rcourtman
|
91dbc94b5f |
Colour pulse-relaxed guests by the thresholds their alerts use (#2735)
The `pulse-relaxed` Proxmox tag lifts a guest's CPU, memory and disk alert triggers to a 95/92/95 floor (applied in `effectiveAlertPolicyNoLock` since #2705), but the browser coloured bars from the configured thresholds alone. A relaxed guest at 88% memory showed a red bar while the alert engine stayed silent until 92%, and the Alerts thresholds row kept showing 80/85/90 with no sign the tag raised them. That is the gap a user described on #863 (comment of 2026-01-16): the tags override the UI settings "but there's no visible indication in the UI". ## What changes - **Bars follow the tag.** `resolveMetricDisplayThresholds` takes the tags the engine reads for a guest and applies the backend's `applyRelaxedGuestThresholds` step before mapping to display thresholds: unset gets the floor, a lower trigger is raised and keeps its clear point (a missing clear is filled five below, as `ensureHysteresisThreshold` does) unless the raised trigger reaches it, and Off stays Off. A relaxed guest now turns red where its alert fires. - **Only Proxmox guests.** `getWorkloadAlertPolicyTags` and `useThresholdsGuestData` pass tags only for Proxmox VMs and LXCs, the only guests whose tags reach alert policy. vSphere and TrueNAS VMs (also `type: vm`) and containers keep their thresholds whatever they are tagged. - **The thresholds row says what is in effect.** A relaxed guest's Alerts → Thresholds row (table and card) adds one line, "Proxmox tag pulse-relaxed: alerts at CPU 95% · Memory 92% · Disk 95%", listing only the thresholds the tag actually raises. It follows unsaved edits. The cells keep the configured values the editor changes. ## Why the row and not the drawer or table Following the user-lens procedure: the job is "I tagged this noisy VM pulse-relaxed: is Pulse honouring it, and when will it alert?". Before this change the bar answered it wrongly at zero clicks, and the only surface that states thresholds (Alerts → Thresholds → Proxmox → VMs & Containers, then search) stated the wrong numbers. No user doc describes the tag, and GitHub comments give conflicting values (95/95/98 in #323, 95/92/95 in #863), so the row has to state the numbers. A badge in the guest drawer or table row would fail the actionability test: once the bar colour is right, the user has nothing to do there, and the place to act is the thresholds row (or the tag in Proxmox). ## Verification - Private mock stack from this branch with a dev-only overlay tagging two demo profiles `pulse-relaxed` (68 VMs). Before the change: relaxed `ci-runner-117` at 88% memory was red. After: every running relaxed row matches the relaxed band (`analytics-batch-70` at 87% reads warning), while untagged guests at 86–92% stay red, at desktop and 375px. The thresholds row line was checked in the card layout, in the table layout (full-width mode at 2560px), at 375px, and through an unsaved edit (Memory 96 drops it from the line). Receipt: `frontend-modern/browser-verification/1f9dcc1c17da6345.json`. - Proof files: `metricThresholds.test.ts` (relaxed resolution, Off, unset, clear filling at margins 5 and 10, scope gating), `workloadTopology.test.ts` (Proxmox-only tags, legacy tag strings), `ResourceTable.test.tsx` (desktop and mobile line, omitted metrics, edits, disabled and untagged rows). `GuestRow.test.tsx` gets the new export in its topology mock. - On pulse-dev: tsc, `npm run lint` and the Alerts, Workloads, utils, stores, platformPage, shared and `src/__tests__` suites. The broad run's only failures were missing repo-root files from a partial sync and 5s timeouts at load ~65; all 15 affected files pass when rerun with the full tree. ## Known limits (existing, not widened here) - The display resolver does not read filter-driven custom rules, which the engine folds in before the override, so a guest a custom rule matches can colour by thresholds the engine does not use. Custom rules have no UI today. The alerts contract now says so. - The Alerts page's guest list memo compares only id and name, so adding or removing the tag in Proxmox shows on that page after a reload. Workloads rows update live. - The thresholds page reads an unset global default as Off (`getTriggerValue`), so the line omits the floor the engine would give a genuinely unset default. A save from the Alerts page always writes these defaults. - `pulse-no-alerts` and `pulse-monitor-only` still have no indication on the row. They are tracked with the rest of the #863 comment in the demand ledger. |
||
|
rcourtman
|
1c6d2a0b2f |
Stop Storage scrolling back to open pools and reopening groups (#2733)
Follow-up to #2709. ## Problem On Proxmox > Storage (pools view): 1. With a pool open, scrolling the page back to the top pulled it back down to the open pool a second or two later, and kept doing so. 2. With `?group=node`, collapsing the open pool's group (Collapse button or header click) bounced it straight back open. ## Causes and fixes - **Bridge reveal re-ran on unrelated state.** `useSummaryTableFocusBridge` called the consumer's `revealActiveSeries` inside a tracked effect, so Storage's reads (the pool-to-group lookup rebuilt on every live update, the expanded group set) re-ran the reveal and its inline-detail scroll. The effect now depends only on a memo of the focused series id and the table root, and runs the callback and scroll untracked (`summaryTableFocus.ts`). Cleanup is registered before the first attempt, so a focus change or unmount cancels pending settle frames. - **Expansion state reopened collapsed groups.** `syncExpandedStorageGroups` re-added every group key missing from the expanded set whenever live data rebuilt the groups, so any collapsed group reopened within ~10 s. The bounce above had been hiding this. `useStorageExpansionState` now stores only the collapsed groups and derives the expanded set, so new groups still open by default. Group keys are bare labels, so a grouping change clears the collapsed set. That stops a label carrying collapse state into another grouping or hiding the headerless `All` group. ## Evidence 12-node mock stack, Chromium at 1440x900 and 390x844, same stack swapped between #2709's code and this change: | Check | Before | After | |---|---|---| | App-shell scrollTop after scrolling away from an open pool (every 0.5 s, 12 s) | 0, 0, 1008, then 1111 ×21 | 0 ×24 | | Collapse open pool's group via button, group state (0.5 s samples) | open ×12 | shut ×32 (16 s) | | Same via header click | open ×12 | shut ×32 (16 s) | | Disks view / Workloads table scroll-away | 0 ×24 | 0 ×24 | Unchanged: lower pool, disk and guest rows open their detail with the row header and detail top in view; Escape closes it; `?resource=ceph-rbd` opens in view; collapsing `shared` and then following an in-app `?group=node&resource=ceph-rbd` link reopens the group and shows the detail; a `?summaryGroup=` deep link scrolls to the pinned header; regrouping node → type → none → node shows every pool. Tests: new bridge tests (`summaryTableFocus.test.tsx`), Storage-level live-update and collapse tests (`Storage.test.tsx`), expansion hook tests, and a guardrail pin that the reveal runs untracked. All six new tests fail on #2709's code. The three expansion tests also fail with the new bridge but the old expansion state. ## Known limit Filter changes do not re-run the reveal. A focused pool whose row stays filtered out past the reveal's ~1.2 s retry window is not revealed when the filter clears. Storage filters live in the URL and a storage link without filter parameters resets them, so only a link carrying a filter that excludes its own pool reaches this. Likewise, with a pool already open and its group collapsed, an in-app link to that same pool leaves the group collapsed, because focus does not change. |
||
|
rcourtman
|
295edef4d5 |
Bind SAML logins to the browser that started them (#2729)
## Why
With `allowIdpInitiated` off, the default, every SAML login failed.
`SAMLService.ProcessResponse` passed crewjam/saml an empty list of
request IDs, and crewjam refuses any Response whose `InResponseTo` is
not in that list, so both SP-initiated and unsolicited Responses ended
in "Authentication failed". Only providers with IdP-initiated logins
enabled, which skips the check, could sign in.
## What changes
- `/api/saml/{id}/login` records each AuthnRequest ID, on the provider's
service, with the login's `returnTo` and the hash of a random token it
sets as the `HttpOnly`, `SameSite=Lax` cookie `pulse_saml_login`
(`__Host-pulse_saml_login` over HTTPS). A browser that already holds a
token keeps it, so logins started one after another in several tabs all
stay valid. Records last ten minutes, 1,024 per provider, oldest
evicted.
- The IdP posts its Response cross-site, and browsers withhold
`SameSite=Lax` cookies from cross-site POSTs. The ACS answers a
cookieless POST with a small page that re-posts the same fields once
from Pulse's own origin (no form action, so path prefixes survive; CSP
nonce on the script; a no-JS Continue button; a marker field stops a
second repost).
- `ProcessResponse` reads the answered request ID from the canonical
content goxmldsig verified for the signature covering the assertion
crewjam returns: the signed Response (or the Response in a signed
ArtifactResponse), otherwise that assertion's own
`SubjectConfirmationData`. It spends the record only when it is bound to
the presenting browser, and the login returns to the recorded `returnTo`
rather than the posted `RelayState`. A Response for another browser
leaves that browser's record intact.
- The signature verifier refuses signed content carrying a prefixed
attribute or namespace declaration named like an un-namespaced attribute
crewjam reads. encoding/xml fills such a field from any attribute with
that local name, and exclusive canonicalization leaves an unused
declaration out of the signed form. On the base commit, appending
`xmlns:IssueInstant` and `xmlns:NotOnOrAfter` to a two-hour-old signed
Response got it accepted, so with `allowIdpInitiated` on a captured
login could be replayed indefinitely.
- RelayState is now URL-escaped into the redirect, so a `returnTo` with
its own query string reaches the IdP whole.
- With `allowIdpInitiated` on, behaviour is otherwise unchanged: no
cookie or repost needed, posted `RelayState` honoured.
- The SLO LogoutRequest store from #2713 is generalised into a shared
record store; its `consume` semantics are unchanged.
Stacked on #2714 (on #2713, on #2707). Those three conflict with main in
`storage-recovery.md`, where their boundary notes and main's land at the
same anchor, so none can merge on its own. This branch carries a merge
of main that keeps both sets of notes; landing it lands the stack.
## Verification
- New tests in `internal/api/saml_handlers_more_test.go`,
`saml_service_test.go` and `audit_handlers_test.go`: default-settings
login through the repost, multiple tabs, refusal of another browser's,
unknown, unsolicited and replayed Responses (the refused browser's own
login still completes afterwards), the `xmlns:InResponseTo` spoof with
the Response signed and with only the Assertion signed, a
confirmation-less signed assertion spliced ahead of another browser's
assertion, assertion-only signed success, stale IdP-initiated Responses
with spoofed ages, repost escaping and CSP nonce, and the audit event
for a binding refusal.
`TestContract_SAMLManualIDPCertificateVerifiesIdPSignatures` now runs at
default settings instead of enabling IdP-initiated logins.
- Negative controls by `-overlay`: with the attribute refusal disabled
the stale-Response replay is accepted again; with the refusal disabled
and the binding read from crewjam's parsed fields the
`xmlns:InResponseTo` spoof logs in; the round-one binding code accepted
the spliced-assertion case as the other user.
- A real Chromium run against a test IdP on `127.0.0.1` and the SP
handlers on `localhost`, behind `SecurityHeadersWithConfig` in
production CSP mode: the IdP's POST arrived `Sec-Fetch-Site: cross-site`
without the login cookie, the repost arrived `same-origin` with it, and
the login landed signed in at
`/alerts?range=7d&saml=success&tab=history`.
## Known limits
- Records live in memory: a restart, a provider re-initialization, or a
login started on a host that does not share cookies with the configured
public URL leaves the Response unmatched, and the user signs in again.
- Two logins started at the same moment in a browser that holds no token
yet each mint one; the last cookie wins and the other login fails.
- As with OIDC login state, a sustained flood of unauthenticated login
starts can evict pending records.
- On plain HTTP, or HTTPS that Pulse does not recognise behind a proxy,
the cookie has no `__Host-` prefix, so a sibling subdomain could plant
one, the same exposure the session cookie has there.
|
||
|
rcourtman
|
8d6f489d3e |
Wrap long identity chips inside phone drawers (#2731)
Below the lg breakpoint, the inline drawer's Identity section rendered alias and IP chips as hand-rolled nowrap spans inside a fixed-layout table cell. Generated aliases such as `proxmox-lxc-docker:Production West:pve1:113` or `k8s-production-1:deployment:redis-ffcd-b17c78f012` are wider than that value cell on a phone, so the cell's hidden overflow and the drawer's clipping content shell cut them off with no ellipsis. The part that distinguishes two aliases is at the end, and a phone has no hover to show the title. - `MetadataBadge` (the contract owner of read-only chip whitespace handling) gains an opt-in `wrap` mode: the chip is capped at its container and breaks inside a token only where no space or hyphen fits. Default stays nowrap; existing consumers get byte-identical classes. - The drawer's IP, alias and alias-count chips in `ResourceDetailSummary.tsx` compose `MetadataBadge`; identity values wrap in full. Desktop drawers are unchanged (one line per alias). - The shared template registry requires the consumer and forbids the old hand-rolled chip classes there; the guardrail and unit tests cover the wrap classes. - `04-mobile.spec.ts` opens a Docker container drawer at phone width and fails if any Identity-section element or text crosses the detail cell, table wrapper or drawer shell (values that end in their own ellipsis excepted). Measured on a private mock estate at 390px: 53 cut chips across 92 Docker/Kubernetes/Proxmox/Machines drawers before, 0 identity chips past the shell across 140 drawers at 390 and 768px after. Rebased on main after #2703 (wider phone drawers): the new spec passes on the Pixel 5 and iPhone 12 projects and fails with the two runtime files reverted (`proxmox-lxc-docker:...:113` still 71px past the cell). Correlation-summary and redaction-label chips were checked and left alone: redaction labels are a fixed five-word vocabulary, and the correlation chips sit outside the nowrap table and break at spaces and hyphens. |
||
|
rcourtman
|
826252d274 |
Stop sending the ignored autonomous_mode chat field (#2730)
## Why
Since
|
||
|
rcourtman
|
2d5ea5758b |
Name sensor setup notice nodes the way the nodes table does (#2727)
The outdated sensor setup notice on the Proxmox page named each node by
its Proxmox node name, while the nodes table directly below it, the
outdated agent notice beside it and Infrastructure settings label the
same node by its display name. With display names set, the notice read
"pve3 is using an older temperature monitoring setup..." above a table
with no row called pve3.
The notice now leads with the label the nodes table shows for the row
and adds the Proxmox node name in parentheses:
> West Production C (pve3) is using an older temperature monitoring
setup that cannot read SATA/SAS disk temperatures.
Both names are needed. The user has to match the notice to a table row,
and the nodes table switches to the Proxmox node name in its narrowest
layout (it showed `pve1`, `pve3` at a 375px viewport and display names
at 414px and up). They also have to find the host to re-run the setup
script on, which is the Proxmox node. Nodes without a display name read
as before ("pve3 is using ...").
The node name is left out only when the label already contains it as a
whole word, ignoring ASCII case, or as a domain's host part (`PVE1`,
`pve3.lan`, `Rack 4 (pve3)`). This is deliberately stricter than
`hasAlternateDisplayName`, which the guest table's group header uses.
That helper normalises punctuation, so it would show `PVE3` alone for
node `pve-3` and hide the host the user has to act on.
The notice's own test file is also registered as proof for the
frontend-primitives policy that covers the notice, matching the sibling
agent notice's test.
Verification:
- Unit tests cover the renamed case, labels that already contain the
node name, labels that only resemble it (`PVE3`/`pve-3`, `pve5-backup`,
`pve10`/`pve1`, a non-ASCII lookalike), blank names and
`nodeName`-before-`node` precedence. The platformPage, SharedPrimitives
guardrail, resource-boundary, nodes and Proxmox page contract suites
pass (396 tests), and eslint, prettier and tsc are clean.
- Browser pass on a private mock stack with three legacy-format nodes,
at 1440, 768, 414 and 375 wide on Overview and Storage. The notice reads
"Affected: Automation Lab Standalone (standalone2), West Production A
(pve1), West Production C (pve3)." with no overflow and no console
errors. Before control with main's two source files: "Affected: pve1,
pve3, standalone2."
|
||
|
rcourtman
|
19b963be71 |
Drop the dead recovery serializer and its stale contract prose (#2728)
The aggregate Recovery page (`frontend-modern/src/components/Recovery/`,
`features/recovery/useRecoverySurfaceState.ts`) was deleted on
2026-05-26 (
|
||
|
rcourtman
|
d1ef487e09 |
Stop hidden tooltips and breakpoints listening to resize (#2724)
On `/proxmox` with 140 workload rows, 859 window `resize` listeners were active after load (203 at phone width with 36 rows), so every window resize ran one callback per listener: - 699 from `useTooltipState.ts` (`useTooltipLayoutState`), which registered a listener for every tooltip instance, including the hidden `TooltipPortal` on each row's metric bars and sparklines. - 156 from `useBreakpoint.ts`, where every call added its own listener and width signal. ## Change - Tooltip layout state listens only while its tooltip is visible and re-reads the viewport when it opens. `useTooltipPortalState` uses the same visibility gate as the portal's render (`when && supportsHoverTooltips()`), so a portal suppressed on a touch device holds no listener either. - `useBreakpoint` keeps one module-level width signal and shared memos, fed by a single rAF-debounced listener while at least one caller is mounted. A new caller syncs to the live `innerWidth`, so a caller created after resizes nobody heard starts correct. ## Evidence Live pass on a private mock-mode stack (headless Chromium, add/removeEventListener wrapped to track live listeners by registering frame): | | before | after | |---|---|---| | 1440×900, 140 rows | 859 active, 859 callbacks per resize | 5 active, 5 callbacks | | 390×844, 36 rows | 203 active | 5 active | - An open row tooltip adds exactly one listener, stays open through window resizes (its viewBox follows 1440×700 and 1380×640), and its listener is removed on close. - Crossing the phone breakpoint in both directions still switches the workload table (`workload-table--desktop` ↔ `--mobile`) and each row's layout. - New tests: `useBreakpoint.test.ts` (one listener for 50 callers, back-to-back resizes leave one pending frame, every caller crosses a breakpoint, a new caller starts from the live width) and `TooltipPortal.test.tsx` (hidden and touch-suppressed portals add no listener, an open portal re-clamps on resize, a portal opened after an unheard resize uses the live viewport). The listener assertions fail against main. - 198 frontend test files / 3620 tests pass across `components/shared`, `components/Workloads`, the hooks, AppLayout, AIChat, Docker tables and the boundary tests; prettier, eslint and tsc are clean. Contract notes: `frontend-primitives.md` (tooltip owner) and `performance-and-scalability.md` (new current-state section). `useBreakpoint.ts` has no owning subsystem, so its rule lives in the performance contract. |
||
|
rcourtman
|
6d8dc0aa73 |
Correct stale platform-page embed contract prose (#2726)
Several v6 subsystem contracts described platform pages embedding
canonical surfaces through controls that no longer exist: a `tableOnly`
/ `embedded tableOnly` mode, a `showFilterToolbar` prop, a
`RecoverySurface`, `forcedPlatformFilter`, an embedded `StorageSurface`
at `/truenas/storage`, and a TrueNAS Protection tab built on a Recovery
surface. The aggregate Recovery surface was deleted on 2026-05-26
(
|
||
|
rcourtman
|
b7bdf1b4eb |
Split a manually linked pair when the API unlinks it (#2725)
## Problem
`POST /api/resources/{id}/unlink` answered 200 "Resources unlinked" but
only recorded a `ResourceExclusion`. The manual link row stayed in
`resource_links`, and `ResourceRegistry.applyManualLinks` never consults
exclusions, so the pair stayed merged in the monitor's registry
rebuilds, the resources API and the websocket broadcast. Repro: link a
standalone agent into a VM, unlink it, then `GET /api/resources/<vm>`:
the agent facet is still there and the agent row is still gone.
Separately, the websocket broadcast coalesced host views with
`CoalescePresentationHostResources`, which applies no exclusions, while
REST's `ListForPresentation` applies the registry's, so a host pair the
operator split could be one broadcast row and two REST rows.
No frontend code calls `/api/resources/{id}/link` or `/unlink`
(`/api/agents/agent/unlink` is a different endpoint).
## Semantics
A resource pair carries one operator decision, and the later one
replaces the earlier: a link joins the pair, an exclusion (unlink,
report-merge) keeps it apart.
- Store writes: `AddExclusion` deletes the pair's link rows and
`AddLink` deletes its exclusion rows, in one transaction, matching both
ID orders; each also deletes a row of its own kind that canonical-ID
succession stored in the other order. The in-memory store does the same
and stamps `CreatedAt` like SQLite.
- Store reads: `GetLinks` / `GetExclusions` keep only each pair's latest
decision (`effectiveManualPairDecisions`), so rows written before this
change or rewritten onto one pair by succession resolve by their later
write; a tie keeps the pair apart.
- Registry: `loadOverrides` resolves its two reads together, so an
unlink committed between them cannot load both decisions into one
generation.
- Identity pins: while a Proxmox node was linked to an agent, the merged
resource's pin could record the node's endpoint hostname with the
agent's machine key. After unlink the node completed that key, minted
the agent's ID and folded into it by ID, a path no exclusion check sees.
A pin no longer completes the identity of a record the operator split
from the pinned resource.
- Broadcast: coalesces through `MonitorAdapter.CoalesceForPresentation`
on the read state that listed the rows when its registry is store-backed
(the store's adapter or a host-continuity overlay), else on the resource
store's adapter (the mock view lists through a store-less registry),
sharing `presentationExclusionFilter` with `ListForPresentation`.
In production REST seeds from the monitor's already-linked listing, so
an unlink shows on REST, `/api/state` and the websocket from the
monitor's next registry rebuild, the same generation in which a new link
reaches the broadcast.
## Known limits (unchanged, documented in the contract)
- Report-merge derives candidate IDs from the merged resource's type, so
it does not name the folded side of a cross-type manual link and cannot
undo one; unlink does.
- An agent's own declared node link (`resolveLinkedResource`) still
ignores exclusions.
- Presentation coalescing compares the IDs of the rows being merged, so
a third same-host fragment that merges first can carry an excluded pair
into one row (REST and broadcast now agree on this).
## Verification
- `TestResourceUnlinkSplitsManuallyLinkedPair` (new
`resources_unlink_test.go`) drives link, unlink and relink through the
API handlers against a `MonitorAdapter` on the same SQLite store,
rebuilt twice per step. It failed on main at "after unlink" exactly as
reported.
- `TestUnlinkSplitsHostPairWhosePinTookTheOtherSidesKey`: node/agent
pairs (IP and FQDN endpoints) through three rebuilds per step; both FQDN
cases fail with the pin guard removed.
- `TestManualPairDecisionReplacesThePreviousOne`,
`TestManualPairDecisionReadsResolveStoredConflictsByRecency`,
`TestRegistryResolvesDecisionsAcrossItsTwoReads`,
`TestMonitorAdapterCoalesceForPresentationHonoursExclusions`,
`TestBroadcastPresentationCoalesceHonoursMergeExclusions`,
`TestBroadcastPresentationCoalescePrefersTheListingReadState`.
- Negative controls (each fails with its fix reverted): store write
rule, newest-link-per-pair, registry combined resolution, pin guard,
broadcast exclusion coalesce, broadcast read-state preference,
store-less adapter fallback.
- Full `./internal/unifiedresources/...`,
`./internal/api/resourceapi/...` and `./internal/monitoring/...` pass.
- Live replay on private mock stacks built from main and from this
branch (`POST .../link`, `.../unlink`, `.../link` on a mock VM and a
standalone agent, then `GET /api/resources/{id}` for both). Main: after
unlink the VM still carries the agent facet and the agent row is still
404, unchanged 25 s later. This branch: after unlink the VM has no agent
facet and the agent row is back at once; relink folds it again.
Report-merge and agent-declared node links are left for separate
follow-ups.
|
||
|
rcourtman
|
ce338b1496 |
Remove the unreachable Workloads group drawer and group pin (#2723)
## Why `useWorkloadsState` defaulted `groupNodeDrawerMode` to `'inline'`, and only that mode wired group-header hover preview, click-to-pin group scope and an inline `NodeDrawer` under the group row. The only page that used the default, the standalone `/workloads` page, was retired in May. Both production mounts have passed `'disabled'` since host drawers moved to each page's own hosts table: Proxmox in |
||
|
rcourtman
|
544588b117 | Merge remote-tracking branch 'origin/main' into fix/drop-unread-circuit-breaker-read | ||
|
rcourtman
|
39b6007f5d |
Pin each manual-link side from its own identity (#2719)
Stacked on #2715 (its commit is this branch's parent; whichever lands first, the other still merges). ## Problem A manual link merges two resources for display, but `PersistIdentityPins` pinned the merged identity. A link primary was pinned with the folded resource's machine key, and the store's one-owner-per-key rule deleted the folded resource's own pin. For a Proxmox node with a qualified endpoint (`https://pve1.example.test:8006`) linked to a host agent, the node's pin records the endpoint, so the next rebuild completed the agent's machine key from it, minted the agent's machine-derived ID and merged the two by ID before the link applied. With the node as primary, succession then rewrote the link into a self-link on the agent's ID; in either direction, removing the link left the pair merged. Reproduced through `MonitorAdapter.PopulateSnapshotAndSupplemental` with a `MemoryStore` on #2715's head. ## Change - `applyManualLinks` captures each side's own pin before the merge (`recordLinkOwnPin`, first capture per rebuild wins), and `PersistIdentityPins` writes those instead of the merged projection, keeping a pinnable folded side pinned under its own canonical ID. - Link sides write through a new `ResourceStore.ReplaceResourceIdentityPins` (both stores). Like an upsert the row keeps fields the side's own pin leaves empty, but it drops a strong key another resource holds this rebuild, which is how a key an earlier release borrowed from the merged identity leaves the row. - A folded side drops a key a live resource or an earlier folded side already holds instead of trading it every rebuild, and writes no pin when that would cost it its machine evidence (a cluster-only row would be open to cluster-slot succession). - Proxmox auto-links (LinkedAgentID / identity matcher) merge before any manual link applies, so their boot-window pins are unchanged (`TestIssue1753*`, `TestMonitorAdapterRebuildPersistsIdentityPins` pass). Known limits, written into the contract: a node row an earlier release already filled with the agent's key still merges the pair by ID while the node can complete from it; two live resources sharing a strong key still alternate, as on main; evidence a later record pass of the same rebuild merges into an already-joined row reaches neither side's pin. ## Verification - New tests in `registry_merge_policy_test.go` (standalone and clustered node, both link directions, three rebuilds with no steady-state pin writes, then unlink separates the pair; shared-key, kept-field, hostname-guard, machine-loss and borrowed-key cases) and `store_test.go` (replace vs upsert on memory and SQLite). - Each new test fails against #2715's runtime files via `go test -overlay`; each guard rule fails its own test when removed. - `internal/unifiedresources` full package, `internal/monitoring` link/identity tests, `internal/api/resourceapi`, `go build ./...` and `go vet` pass. - Contract: `unified-resources.md` clause 25 and "Operator links reach record-ingested resources". |
||
|
rcourtman
|
bb35a5c006 |
Correct stale Storage embed and History range contract prose (#2718)
Three contract passages described toolbar controls that no longer exist. **storage-recovery.md (StorageSurface embed).** The contract said platform pages opt into the Storage toolbar through a `showFilterToolbar` prop on `StorageProps`, keep `tableOnly` to hide a storage summary section, and get a chart-collapse control. It also mentioned `RecoverySurface`, `forcedPlatformFilter`, and an embedded `StorageSurface` at `/truenas/storage`. On main: - `Storage.tsx` always renders `StoragePageControls`; only kiosk mode hides it. - There is no summary section and no chart-collapse control. - The Proxmox Storage tab is the only embed. It passes `forcedSourceFilter`, `resourceSource`, `suppressNodeFilter`, `filterAriaLabel` and search copy. - `/truenas/storage` renders `TrueNASStorageTopologyTable`. The rewrite describes what `Storage.tsx`, `StoragePageControls.tsx` and `useStoragePageFilters.ts` actually do: - The forced source is read as the route source and never written back to the URL. - The Source chip is auto-suppressed and does not surface the clear action. - `suppressNodeFilter` drops the Node chip and resets a selected node to all nodes. **performance-and-scalability.md and frontend-primitives.md (Workloads History range).** Both said the sparkline range shows only while Trends is active and carries a `Trend range` label. `WorkloadsFilter` actually shows it in Trends, and in Bars at desktop widths while Row hover is `History`. Its label is `History`. **Test.** A new rendered test in `Storage.test.tsx` pins the embed behaviour the rewrite claims. Five negative controls each fail it: - dropping the forced-source suppression - dropping the node reset - unlocking the route source - gating the toolbar off - keeping the Node filter The `History` label was already pinned by `WorkloadsFilter.test.tsx` and `WorkloadsSurface.performance.contract.test.tsx`. Another stale test title, "table-only embedding", is renamed. Sibling stale passages are left for a follow-up: `embedded tableOnly` in frontend-primitives.md and cloud-paid.md, the WorkloadsSurface `showFilterToolbar` text in performance-and-scalability.md, and the TrueNAS StorageSurface line in unified-resources.md. |
||
|
rcourtman
|
1b01ab6e1b | Merge branch 'refactor/drop-unread-patrol-state-members' into fix/drop-unread-circuit-breaker-read | ||
|
rcourtman
|
aac5ca5fc2 |
Colour pulse-relaxed guests by the thresholds their alerts use
The pulse-relaxed Proxmox tag lifts a guest's CPU, memory and disk alert triggers to a 95/92/95 floor, but the browser coloured bars from the configured thresholds alone. A relaxed guest at 88% memory showed a red bar while the alert engine stayed silent until 92%, and its Alerts thresholds row kept showing 80/85/90 with no sign that the tag raised them (issue #863, comment of 2026-01-16). - resolveMetricDisplayThresholds takes the tags the engine reads for a guest and applies the backend's applyRelaxedGuestThresholds step: unset gets the floor, a lower trigger is raised and keeps its clear point unless the raised trigger reaches it, and Off stays Off. - getWorkloadAlertPolicyTags passes tags only for Proxmox VMs and LXCs, the only guests whose tags reach alert policy. vSphere and TrueNAS VMs and containers keep their thresholds whatever they are tagged. - A relaxed guest's Alerts thresholds row and phone card add one line naming the tag and the thresholds it raises. It follows unsaved edits, while the cells keep the configured values the editor changes. |
||
|
rcourtman
|
32fad38aa4 | Merge remote-tracking branch 'origin/main' into refactor/drop-unread-patrol-state-members | ||
|
rcourtman
|
199f45e211 | Merge origin/main into refactor/drop-chart-visibility-toggle | ||
|
rcourtman
|
04f61d431f |
Remove the unreachable Workloads chart visibility toggle
WorkloadsFilter rendered ChartVisibilityToggleButton only when a surface passed onChartsToggle, and none has since the workload summary chart section was deleted with the summary components. The button had no other production consumer, so the shared primitive, its two shared-template registry entries, the filter-model props and the tests that pinned them described a Show/Hide charts control that nobody could reach. Delete them. The guardrail that required the primitive now keeps the exact retired registry ids, export and props out: a returning summary section is a product decision that should register its own control. Contract prose in frontend-primitives, performance-and-scalability and unified-resources now says no chart show/hide control exists instead of describing the deleted one. No visible change: the Workloads View disclosure on the Proxmox and vSphere overviews shows the same sections and controls before and after, at desktop and phone widths. |
||
|
rcourtman
|
4e95ca45cb |
Stop Storage scrolling back to open pools and reopening groups
On the Storage pools view, an open pool pulled the page back to itself a second or two after the operator scrolled away, and collapsing the open pool's group bounced it straight back open. The summary focus bridge ran the consumer's revealActiveSeries callback inside a tracked effect, so everything Storage's callback reads (the pool-to-group lookup that every live update rebuilds, the expanded group set) re-ran the reveal and its inline-detail scroll. The reveal now subscribes only to the focused series id and the table root, and runs the callback and scroll untracked, so it runs once per deliberate focus change. Its cleanup is registered up front, so a focus change or unmount also cancels pending settle frames. Fixing that exposed a second cause of the bounce: the expansion state re-added every group key missing from its expanded set whenever a live update rebuilt the groups, so any collapsed group reopened within about ten seconds. It now stores only the groups the operator collapsed and derives the expanded set, so new groups still open by default. Group keys are bare labels, so a grouping change clears the collapsed set rather than carrying collapse state into another grouping or hiding the headerless "All" group. Measured on a 12-node mock stack: after scrolling away from an open pool the shell stays at scrollTop 0 for 12 s (was 0 -> 1111 within ~1.5 s), and a group collapsed by its button or header stays collapsed for 16 s. The new bridge, Storage and expansion tests fail on the previous code. Filter changes do not re-run the reveal either, so a focused pool whose row stays filtered out past the reveal's retry window is not revealed when the filter clears; storage links reset the URL-owned filters, so only a link that carries an excluding filter gets there. |
||
|
rcourtman
|
1eb4a84fe6 |
Drop the dead recovery serializer and its stale contract prose
The aggregate Recovery page and its route-state owner were deleted on 2026-05-26, but RECOVERY_QUERY_PARAMS, parseRecoveryLinkSearch and buildRecoveryRouteSearch stayed in resourceLinks.ts with no production caller, which kept recoveryItemTypePresentation.ts alive as their only reader. Both go. The frontend-primitives, storage-recovery, unified-resources, alerts and cloud-paid contracts still described the deleted page as live: its shell, event-first workspace, section owners, filter and column models, recovery route state, and the incident panel's link chips. Those passages now say what exists. Recovery rows render in the Proxmox Backups and TrueNAS Protection tabs, useRecoveryPoints decodes every recovery-points read, and rules that only governed the deleted page are gone or reduced to a reintroduction guard. New assertions pin the rewritten claims: resourceLinks exports nothing recovery-shaped, the recovery-points hook hands out normalized points and is the only reader, the retired hooks stay deleted, each tab keeps its own view and filter state, and the incident panel imports no route helpers. |
||
|
rcourtman
|
d95a316118 |
Stop sending the ignored autonomous_mode chat field
Since
|
||
|
rcourtman
|
59cc3ce10b |
Wrap long identity chips inside phone drawers
Below the lg breakpoint, the inline drawer's Identity section rendered alias and IP chips as hand-rolled nowrap spans inside a fixed-layout table cell. A generated alias such as proxmox-lxc-docker:Production West:pve1:113 or k8s-production-1:deployment:redis-ffcd-b17c78f012 is wider than that value cell on a phone, so the cell's hidden overflow and the drawer's clipping content shell cut it off with no ellipsis. On a 390px mock estate 53 chips across Docker and Kubernetes drawers were cut. The part that tells two aliases apart sits at the end, and a phone has no hover to show the title. MetadataBadge owns read-only chip whitespace handling, so it gains an opt-in wrap mode that caps the chip at its container and breaks inside a token only where no space or hyphen fits. The drawer's IP, alias and alias-count chips now compose MetadataBadge, and the identity values wrap in full. The default stays nowrap, so existing badge consumers are unchanged, and desktop drawers still show each alias on one line. The shared template registry now requires ResourceDetailSummary to compose MetadataBadge and forbids the old hand-rolled chip classes there. A mobile integration check opens a Docker container drawer and fails if any Identity-section element or text crosses the detail cell, table wrapper or drawer shell; it fails on the old code and passes with this change. |
||
|
rcourtman
|
077c3a8959 |
Merge origin/main into fix/saml-acs-request-binding
Brings the branch up to the main it lands on; the merge is clean. Contract-Neutral: Merge of origin/main into the SAML stack: main's changes were governed when they landed, and the merge is clean |
||
|
rcourtman
|
d72d355ccf |
Merge origin/main into fix/saml-acs-request-binding
The SAML stack this branch builds on appends its boundary notes to storage-recovery.md at the anchor main used for the linked-agent stale flag note. Both sets of notes are kept, the SAML ones first. Contract-Neutral: Merge of origin/main into the SAML stack: main's changes were governed when they landed, and the only resolution keeps both sets of storage-recovery boundary notes |
||
|
rcourtman
|
4d4e2e0159 | Merge remote-tracking branch 'origin/main' into fix/sensor-setup-notice-names | ||
|
rcourtman
|
371435e1eb |
Name sensor setup notice nodes the way the nodes table does
The outdated sensor setup notice named each node by its Proxmox node
name ("pve3") while the nodes table below it, the outdated agent notice
beside it and Infrastructure settings label the same node by its
display name ("West Production C"). A user with display names set could
not tell which row the notice meant.
The notice now leads with the label the nodes table shows for the row
(the resource name, else the id) and adds the Proxmox node name in
parentheses: "West Production C (pve3)". Both names are needed. The
nodes table switches to the Proxmox node name in its narrowest layout,
and the setup script is re-run on that node, so the exact native name is
what the user needs to find the host.
The node name is left out only when the label already contains it as a
whole word, ignoring ASCII case, or as a domain's host part ("PVE1",
"pve3.lan"). This is deliberately stricter than hasAlternateDisplayName
in the guest table's group header, which normalises punctuation and would
show "PVE3" alone for node "pve-3", hiding the host to act on.
The notice's own test file was never registered as proof for the
frontend-primitives policy that covers the notice, unlike the sibling
agent notice's test, so it is added to that policy's exact_files.
|
||
|
rcourtman
|
4cf1c318fb |
Correct stale platform-page embed contract prose
Several v6 contracts still described platform pages embedding canonical surfaces through controls that no longer exist: a `tableOnly` / `embedded tableOnly` mode, a `showFilterToolbar` prop, a `RecoverySurface`, an embedded `StorageSurface` at /truenas/storage, and a TrueNAS Protection tab built on a Recovery surface. The aggregate Recovery surface was deleted on 2026-05-26 and neither WorkloadsSurface nor Storage has an embed mode. The platform-page rule in frontend-primitives.md (and its cloud-paid counterpart) now says what the pages do: the Proxmox and vSphere Overview tabs embed WorkloadsSurface with the page owning the workloads state and the one WorkloadsFilter toolbar, the Proxmox Storage tab embeds StorageSurface with forcedSourceFilter, and every other tab, including /truenas/storage, Proxmox Backups and TrueNAS Protection, renders a platform-owned table on the shared platform-page primitives. The old rule said pages only embed canonical surfaces and must not add per-family tables; that has not been true since the platform-first migration, so the rule now names the three embeds, keeps those tabs' rows on them, and keeps the bans on estate summary panels and forked table primitives. cloud-paid defers to that rule instead of carrying a second copy. The new guardrail case pins those embeds per page, the absence of an embed-mode prop on both surfaces, the platform lock on the Workloads embeds, and the recovery tabs' platform-owned tables, so the contract and the code cannot drift apart silently again. |
||
|
rcourtman
|
58ee64fc1c |
Split a manually linked pair when the API unlinks it
POST /api/resources/{id}/unlink recorded only an exclusion. The link row
stayed, and applyManualLinks never consults exclusions, so the monitor's
rebuilds, the resources API and the websocket kept the pair folded while
the API answered "Resources unlinked".
A pair now carries one operator decision. AddExclusion deletes the pair's
links and AddLink its exclusions, in one transaction and in either ID
order. Reads keep each pair's latest decision, so rows written before
this change, or rewritten onto one pair by canonical-ID succession,
resolve by their later write, and a registry resolves its two reads
together so an unlink landing between them cannot load both.
While a Proxmox node was linked to an agent, the merged resource's
identity pin could take the agent's machine key. After unlink the node
completed that key, minted the agent's ID and folded into it by ID, where
no exclusion check runs. A pin no longer completes a record the operator
split from the pinned resource.
The broadcast coalesced host views without exclusions while REST's
ListForPresentation honoured them, so a split host pair was one websocket
row and two REST rows. The broadcast now coalesces through the read state
that listed the rows, with the same exclusion filter.
|
||
|
rcourtman
|
9daa8154c8 | Merge remote-tracking branch 'origin/main' into perf/shared-resize-listeners | ||
|
rcourtman
|
03e4920460 |
Stop hidden tooltips and breakpoints listening to resize
On /proxmox with 140 workload rows, 859 window resize listeners were active after load (203 at phone width with 36 rows), so every window resize ran one callback per listener and each set signals. 699 came from tooltip layout state registering a listener for every instance, including the hidden TooltipPortal on each metric bar and sparkline, and 156 from every useBreakpoint() call adding its own listener. Tooltip layout state now listens only while its tooltip is visible (a portal suppressed on a touch device counts as hidden) and re-reads the viewport when it opens. useBreakpoint keeps one module-level width signal fed by a single rAF-debounced listener while any caller is mounted, and a new caller syncs to the live innerWidth. The page now holds 5 listeners at both widths. An open tooltip adds one and removes it on close, still follows window resizes, and crossing the phone breakpoint still switches the workload table and row layout. |
||
|
rcourtman
|
03767f1e95 |
Remove the unreachable Workloads group drawer and group pin
useWorkloadsState defaulted groupNodeDrawerMode to 'inline', but only the standalone /workloads page used that default, and it was retired in May. Both production mounts, the Proxmox and vSphere overviews, have passed 'disabled' since host drawers moved to each page's own hosts table ( |
||
|
rcourtman
|
1a094646d7 |
Let only POST narrate single-resource reports with Pulse Assistant (#2716)
## What
Single-resource report generation (`/api/admin/reports/generate`) now
takes `POST` with a JSON body as its canonical transport, and only
`POST` may narrate a PDF with Pulse Assistant.
- **POST**: the same fields as before (`resourceType`, `resourceId`,
`format`, `metricType`, `start`, `end`, `title`) in a JSON body, decoded
with the `generate-multi` rules (1MB cap, unknown fields and trailing
payload rejected). When Pulse Assistant is configured, a PDF's executive
summary is narrated by the AI provider as before.
- **GET**: still accepts the same fields as query parameters and returns
the same report, but always uses the deterministic summary. The PDF
shows a short note in place of the "Configure Pulse Assistant" tip,
which would be wrong for a tenant that already has the Assistant
configured.
- **GET never resolves the AI service at all.** It reads Patrol findings
only from a tenant AI service that is already running, through a new
`AISettingsHandler.ExistingAIService` lookup. Constructing a tenant
service through `GetAIService` can list provider models or start
background discovery, so a report GET must not trigger it. The router's
reporting wiring moves into `Router.wireReportingAIResolvers` so a test
can drive it.
- **Other methods**: `405` with `Allow: GET, POST`.
- The shared report body decoder now also returns `body_too_large`,
rather than `invalid_body`, when the cap is hit while reading past a
valid value.
- The reporting catalog advertises
`performanceReport.singleResourceMethod: "POST"`, and the settings UI
follows it. A catalog without the field, and the legacy fallback
catalog, keep the GET query transport that older backends accept. Fleet
reports were already `POST`.
- `docs/API.md` and `docs/MSP.md` (and their shipped copies) document
both methods.
## Why
Before this change, a GET for a PDF resolved the tenant's AI narrator.
That made a paid provider call and recorded a usage event in the AI cost
ledger. GET passes the demo-mode guard and the CSRF check, and a
SameSite=Lax session cookie is sent on cross-site top-level GET
navigations. So following a link could spend a signed-in admin's AI
budget.
GET stays available rather than returning `405` because the API
reference and the MSP guide documented it for scripted exports. Breaking
those scripts would add no protection: CSV is never narrated, and GET no
longer reaches the provider.
## Verification
- `TestContract_SingleReportAINarrationRequiresPOST` drives a real
report engine with a counting narrator. GET, HEAD, PUT and DELETE make
zero narrator calls and POST makes one. The GET narrative is heuristic
and carries the note. As a negative control, the same test fails against
the handler on `main` via `go test -overlay` ("GET report called the AI
narrator 1 times, want 0").
- `TestContract_SingleReportGETDoesNotConstructTenantAIService` drives
the router's real resolvers with an uncached tenant: GET constructs no
tenant AI service, while POST resolves one. A negative control that
swaps `ExistingAIService` back to `GetAIService` fails it ("GET report
constructed 1 tenant AI services, want 0").
`TestAISettingsHandler_ExistingAIServiceNeverConstructs` covers the
lookup itself.
- Handler tests cover the POST body rules (empty, unknown field,
trailing value, form-encoded body, fields only in the query, oversized
padding) and the `Allow` header. The catalog contract snapshot includes
the new field.
- Frontend tests cover POST bodies, the GET transport for catalogs that
name GET, parser defaults and validation, and the legacy fallback.
`docsLinks.test.ts` ties the docs to the handler.
- Browser pass on a private stack: real login, dev Pro licence, and
Pulse Assistant pointed at a local stub OpenAI-compatible provider that
logs every call. Checked at 1280×800 and 390×844:
- Generating a single-resource PDF from Settings → Data & Reports sends
one `POST` with a JSON body, downloads the PDF and calls the provider
once. The downloaded PDF carries the AI narrative.
- CSV makes no provider call. The fleet report still posts to
`generate-multi`.
- A top-level browser navigation to the GET URL with the session cookie
downloads the deterministic PDF, with the new note, and makes no
provider call.
- HEAD, PUT, DELETE and PATCH return `405`, and a session POST without
the CSRF header returns `403`.
- The AI usage summary lists `report_narrative` only for the UI POSTs.
- Full `./internal/api/...` and `./pkg/reporting` run on the dev VM:
everything passed apart from two tests that fail there for environment
reasons in code this change does not touch
(`TestServerInfoEndpointReportsDevelopment` needs `.git`; a `configapi`
shell-hangup bootstrap test).
The same class of issue on the Assistant session routes
(`/api/ai/sessions/{id}/summarize` and friends) is handled separately in
#2670.
|
||
|
rcourtman
|
1a48d27aa0 | Merge remote-tracking branch 'origin/main' into refactor/drop-unread-patrol-state-members | ||
|
rcourtman
|
fd36869223 |
Judge source merges by the ingest's stale thresholds (#2717)
Stacked on #2715 (its commit is the parent of this one); either can land first. ## Problem `mergeInto` passed nil thresholds to `mergeMetrics`, so every source merge into an existing row (snapshot and record ingest alike) judged metric freshness by the registry defaults, while the stale pass and manual links used the monitor's configured thresholds. Two visible results: - A Proxmox node polled every two minutes (four-minute threshold) merged with its auto-linked agent showed the silent agent's CPU and memory, while the stale pass called the node's poll current. - A supplemental refresh (`MonitorAdapter.PopulateSupplementalRecords`, and the read-state overlay) of a linked vSphere VM kept the in-guest agent's memory over vSphere's current reading, where a full rebuild from the same records shows vSphere's. ## Change - Snapshot and record ingest hold the caller's thresholds on the registry while they run (the same per-pass pattern as the guest loop's agent index), and `mergeInto` passes them to `mergeMetrics`. The snapshot helpers that reach `mergeInto` keep their signatures. - The stale pass and the metric merge share one threshold rule, `sourceStaleThreshold`, so the merge also honours a sighting's expected update interval (slow Proxmox disk inventory), as the stale pass already did. - Outside mock mode the monitor configures only Proxmox, PBS and PMG, so those are the production sources this changes; mock mode also configures TrueNAS, vSphere and availability. ## Proof - `TestMonitorAdapterSourceMergesUseConfiguredStaleThresholds`: rebuild, supplemental refresh and overlay of a linked vSphere VM, plus a node with an auto-linked agent, through `NewMonitorAdapterWithStaleThresholds`. - `TestSourceMergesJudgeFreshnessByTheIngestStaleThresholds`: snapshot node+agent merge with and without thresholds. - `TestMetricMergeFreshnessMatchesTheStalePass`: the merge and the stale pass agree for default, configured and expected-interval sightings. - Negative control with #2715's `registry.go`: the supplemental, overlay, node+agent, snapshot and expected-interval cases fail; the rebuild case and the defaults case pass on both. - `go test ./internal/unifiedresources/... ./internal/monitoring/... ./internal/api/resourceapi/... ./internal/mock/...` on pulse-dev. |
||
|
rcourtman
|
7f45148d15 |
Drop the unused Patrol readiness-failure object behind the toggle toast (#2687)
Stacked on #2686 (which stacks on #2682 and #2674); this branch carries
their commits until they land. Only the top two commits are new.
## What
`buildPatrolSettingsReadinessFailure` built a full
`PatrolConfigurationFailureInput` for the Patrol page's inline "needs
attention" state. That state lost its render in
|
||
|
rcourtman
|
796a70999b |
Stop tenant incident stores racing on shared ID counters (#2711)
Each tenant monitor owns its own `IncidentStore`, but incident and incident-event IDs came from package-level `int64` counters bumped with `++` under nothing but the calling store's mutex. Two tenants recording alerts at once raced on them. `go test -race` failed `TestLeavingMockModeClearsEveryRunningTenantMonitor` on #2698 ([run](https://github.com/rcourtman/Pulse/actions/runs/37652317098/job/112919080928)) at `incidents.go:1753`. The remediation record counter in `remediation.go` had the same shape, and `startPatrolForContext` creates one remediation log per org. **Fix** - All six `internal/ai/memory` ID generators (incident, incident event, remediation record, memory, incident memory, pattern memory) now mint through one atomic helper, `nextSequencedID` in `ids.go`. - The numeric suffix no longer wraps at 1000. Even with atomic counters, a generator that minted more than a thousand IDs in one wall-clock second repeated an ID. That matters because `RecordNote` finds an incident by ID and the incident query deduplicates merged events by ID. The `<prefix><YYYYMMDDhhmmss>-<n>` layout is unchanged. Only `n` can now exceed three digits. - `intToString` lost its last runtime callers and is removed. **Proof** - `internal/ai/memory/ids_test.go` mints 2,400 IDs from every generator on 16 goroutines and records alerts into two tenant stores concurrently. - Against the `origin/main` runtime files (via `-overlay`), `-race` reports DATA RACE at `incidents.go:1753/1760` and `remediation.go:355`, and every generator fails the uniqueness/shape check: four produce duplicate IDs, and `mem`/`pat-mem` produce a wrapped `-0` suffix. - With the fix, `go test -race ./internal/ai/memory/` passes, and `go test -race -count=5 -run TestLeavingMockModeClearsEveryRunningTenantMonitor ./internal/monitoring/` passes. **Residual:** across restarts, only the timestamp separates IDs. If the local clock repeats a second (DST fall-back or a backward step), a restarted sequence could still meet an earlier suffix. Persisted IDs and the stored format are unchanged. The `ai-runtime` contract records this. |
||
|
rcourtman
|
1001eb6a69 | Merge remote-tracking branch 'origin/main' into refactor/drop-unread-patrol-state-members | ||
|
rcourtman
|
df885e1f83 |
Take main's removal of two frontend-primitives items
Main dropped items 19 and 20 (summary-card emphasis and density-map overview) in #2709. They sat directly below the FindingsPanel handoff sentence this branch appends to item 18, so the branch stopped merging. Take the same deletion here and move the sentence two lines up inside item 18, so unchanged lines separate it from main's hunk and origin/main merges cleanly. |
||
|
rcourtman
|
89a05ed735 |
Pin each manual-link side from its own identity
A manual link merges two resources for display, and the durable identity pin was written from that merged identity. A link primary was pinned with the folded resource's machine key, and the store's one-owner-per-key rule then deleted the folded resource's own pin. A Proxmox node's pin records its qualified endpoint as the hostname, so on the next rebuild the node completed the linked agent's machine key from that pin, minted the agent's machine-derived ID and merged with the agent by ID before the link applied. With the node as primary, succession then rewrote the link into a self-link on the agent's ID; in either direction, removing the link left the pair merged. The registry now captures each link side's own pin when the link first joins it in a rebuild, and PersistIdentityPins writes those instead of the merged projection, keeping a pinnable folded side under its own canonical ID. Proxmox auto-links merge by identity before any manual link applies, so their boot-window pins are unchanged. A link side rewrites its whole row through a new ReplaceResourceIdentityPins. Like an upsert it keeps stored fields its own pin leaves empty, but it drops a strong key another resource holds, since an earlier release borrowed that key from the merged identity and the store's one-owner rule would otherwise delete one row or the other. A folded side drops a strong key a live resource or an earlier folded side already holds, so the two never trade it on every rebuild, and writes no pin when that would leave it without its machine evidence. A node row an earlier release filled with the agent's machine key is handed back only when the node cannot complete from it. With a qualified endpoint or a cluster name, completion still supplies the borrowed key before the link applies, so such a pair stays merged. |
||
|
rcourtman
|
c5416b6800 |
Alert on the Unraid temperature an agent disk shows (#2701)
Stacked on #2698 (which sits on #2694, #2685 and #2679); this PR's own change is the top commit, "Alert on the Unraid temperature an agent disk shows", plus a merge of main for an alerts.md anchor both touched. ## Why A host agent's physical disk can show a temperature that only the host's Unraid inventory reports: - a SMART row whose smartctl probe returned no temperature takes its Unraid row's reading (`HostSMARTDiskTemperature`), and - a disk only the inventory lists, such as a `--disk-exclude` match, is built from its Unraid row alone (`HostUnraidDiskTemperature`). Physical Disks (Temp cell and the `Running Hot` verdict, "at or above its alert threshold") and Patrol judge that reading against the alert disk temperature policy, but `CheckHost` only read the SMART row's own field. A disk could read as at its alert threshold while no alert fired. Unraid's own warning/critical disk temperature notifications do not make this redundant: Pulse already alerts on the smartctl readings and array state of the same hosts, which Unraid also notifies on. ## What changes - `unifiedresources.HostDiskTemperatureReadings(host)` lists the reading each of a host's disks shows, built with the registry's own adapters: one per SMART row the registry ingests, then one per Unraid row whose disk key no SMART row carries (the same rule as the history writer in #2698). - `CheckHost` judges those readings: - SMART rows keep their `agent:<host>/disk_temp:<device>` resource IDs. An inventory-only disk uses its block device token, which is the ID a plain Linux SMART row for that device uses, so a later SMART reading continues the same alert. - Per-type thresholds and host overrides use the disk type the disk shows. - Only a reading collected now fires or resolves, so a spun-down disk holds its alert. - When conflicting serials put two disks on one device, its alert judges the reading that stands highest against its own type's threshold. - An inventory-only disk matching the agent's `--disk-exclude` raises nothing, and an alert it raised before the exclusion resolves on the next report. This matches the documented exclusion from monitoring and how linked Proxmox hosts already treat excluded disks. - Absence cleanup still runs only on a non-empty SMART list. ## Verification - New alert tests in `internal/alerts/host_unraid_lifecycle_test.go` cover: the fallback, an inventory-only disk, an empty SMART list, two disks on one device (also of different types), a controller member under another device label, standby, exclusion, disk-type and host-override thresholds, holds, handover to a SMART reading, departure, and the exclusion clear. - Each fails on the previous `host.go`. - Removing any guard (exclusion skip, exclusion clear, disk type, collected-now, threshold ranking) fails its own test. - `TestHostDiskTemperatureReadingsMatchTheDisksTheRegistryShows` compares every reading with the disk the live registry ingests, including conflicting serials, a virtual device and a MegaRAID member Unraid lists under its block device. - `go test ./internal/alerts/... ./internal/unifiedresources/...` passes on the merged tree, as do the Unraid/SMART/temperature tests in `internal/monitoring`. |
||
|
rcourtman
|
a14584d9f1 |
Drop the summary focus bridge members nothing reads (#2709)
The jump-to-row affordance lost its only consumer when the summary components were deleted, yet the bridge kept computing it on every Workloads and Storage page: a capture-phase window scroll listener and a resize listener re-measured the hovered or focused row on each frame and hover change, walking its ancestors with getComputedStyle. On the 12-node mock storage table that was 128 calls across eight row hovers and 320 across ten scroll steps, now none. Storage's reveal callback kept branches that switched between the pools and disks views for that jump. Focus only ever names a row in the active view, so drop them and the disk-id memo that served them. The chart-hover input went the same way. No summary chart renders, so setChartHoverSync was never called and the resolver's chart branch was always fed null. Remove the signal, its type and the resolver option, and say in the contract how a returning chart re-enters the shared scope instead. Focus reveal, pinned-group reveal, Escape and row emphasis behave as before at desktop and phone widths; the browser receipt records the before/after pass. |
||
|
rcourtman
|
b3b60f0632 |
Bind SAML logins to the browser that started them
With allowIdpInitiated off, the default, every SAML login failed: ProcessResponse passed crewjam/saml no request IDs, and crewjam refuses a Response whose InResponseTo is not among them. Only providers that allowed IdP-initiated logins, which skips the check, could sign in. MakeAuthRequest now records each AuthnRequest ID with the login's returnTo and the hash of a random token that /login sets as the HttpOnly SameSite=Lax cookie pulse_saml_login (__Host- over HTTPS), reused across tabs. The IdP posts its Response cross-site, which withholds Lax cookies, so the ACS answers a cookieless POST with a page that re-posts the same fields once from Pulse's origin. The Response must answer an unexpired request bound to the presenting browser; it spends that request and returns to the recorded path. Another browser's request is left for its owner. The request ID is read from the signature covering the assertion crewjam returns (the signed Response, else that assertion's own SubjectConfirmationData), not from crewjam's parsed fields and not from other signed assertions riding in the same unsigned Response. encoding/xml fills an un-namespaced attribute field from a namespace declaration with the same local name, and exclusive canonicalization drops an unused one from the signed form, so xmlns:InResponseTo could rebind a signed Response and xmlns:IssueInstant / xmlns:NotOnOrAfter let a stale one pass crewjam's age checks: an IdP-initiated login could be replayed indefinitely. The signature verifier now refuses signed content declaring a prefix named like an attribute crewjam reads, in both modes. RelayState is now escaped into the redirect query, so a returnTo with its own query string reaches the IdP whole. |
||
|
rcourtman
|
e7cabc72f1 |
Hide pprof routes on demo-mode instances (#2708)
The `/debug/pprof` handlers are gated on `RequireAdmin` plus a `settings:read` scope that session logins skip. A `DEMO_MODE` instance that signs visitors in as the configured admin therefore served every visitor goroutine and heap dumps, the process command line, and CPU profiles and execution traces that load the whole process. Demo mode already hides diagnostics, logs, admin users and manual discovery through the admin-operations policy; pprof was missing from it. ## Change - `internal/api/demo_mode_operations.go`: through the admin-operations policy, the demo guard answers `/debug/pprof` and every path below it with `404` for every method, ahead of route authentication, so a write probe such as the `POST` symbol lookup gets the same `404` instead of the generic demo `403`. Instances outside demo mode keep the routes, and `PULSE_PPROF_DISABLED` still removes them everywhere. - Contract notes in api-contracts (owner), agent-lifecycle and storage-recovery. ## Proof - `TestContract_DemoModeHidesPprofFromDemoAdminSession` signs in as the demo admin through the full router. The same session gets `200` from `/debug/pprof/cmdline` below the demo guard, and `404` for eight pprof paths × `GET`/`HEAD`/`OPTIONS`/`POST` through the served handler. Against the previous policy (`go test -overlay`) `GET` and `HEAD` returned pprof output (`200`, or the root's `301` into the index), `OPTIONS` got the router's preflight `200` and `POST` the generic demo `403`. - `TestPublicDemoAdminOperationsPolicyHidesEveryDebugRoute` parses the router files and fails if a literal route under `/debug/`, with or without a method prefix, escapes the policy for any of seven methods, so a new debug handler there cannot ship reachable on a demo by default. A temporary `GET /debug/vars` registration made it fail as expected. - Path-handling probe (not committed): 210 raw requests (35 request targets × six methods), covering trailing and double slashes, dot segments, `%2F`/`%2e` encodings, absolute-form targets and `CONNECT`, run against a copy of the router's mux dispatch. None reached a pprof handler; paths the mux would clean onto pprof only receive a redirect, and the policy catches the follow-up request. |
||
|
rcourtman
|
6263c580f7 |
Judge a linked agent's guest disks by that agent everywhere (#2712)
Follow-up to #2669, which made the guest drawer's filesystem rows follow a linked Pulse agent's freshness. The workload table's disk cell and the drawer History still judged the same guest by the Proxmox read reason, so one guest disagreed with itself across the three. ## What changes - **One rule for three readers.** `getWorkloadGuestDiskRead` (`workloadGuestPresentation.ts`) now drives the table disk cell, the drawer's filesystem rows and History. When the linked agent's filesystems fill in for a VM that has no Proxmox guest filesystems, the reading is current while the agent reports, last known once it is stale, and unavailable once the guest is stopped. Proxmox's own guest filesystems keep their read reason. - **Where the aggregate disk metric comes from.** On a manually linked VM, Proxmox records no guest disk usage without guest filesystems, so the registry's link merge keeps the agent's disk metric. A registry test pins that. - **Poller carry-forward.** `previousVMFromView` built the previous VM from the merged metric, so the next poll carried the agent's disk forward as a Proxmox `prev-agent-not-running` read. Proxmox priority then froze it over the agent's live value. Only Proxmox's own disk reading is carried now (`VMView.DiskFromLinkedAgent`). - **`agent.stale` is always on the wire.** The browser merges agent facets field by field, so an omitted false kept a resumed agent "stopped reporting" until a reload. This affected Machines rows and node drawers too. - **Withdrawn filesystems.** The websocket merge kept filesystems that a native Proxmox VM facet or a native agent facet (which always carries `stale`) no longer reported, including in the `platformData.agent` mirror that host threshold rows read. ## Verification - New tests, each failing on main: `GuestRow.test.tsx`, `GuestDrawer.test.tsx`, `workloadGuestPresentation.test.ts`, `resourceStateAdapters.test.ts`, `TestPreviousVMFromViewKeepsLinkedAgentDiskOutOfProxmoxCarry`, `TestAgentDataAlwaysSendsStale`, `TestVMViewDiskFromLinkedAgentFollowsSelectedMetricSource`, and `TestResourceRegistry_ManualGuestAgentLinkDiskMetricFollowsProxmoxGuestFilesystems`. The registry test is a characterization test; it fails when the agent outranks Proxmox on guest disk. - Private mock stack at 1440 and 390, with agents linked into VMs through `POST /api/resources/{vm}/link`: - Live agent, no QEMU guest agent: before, "- Unavailable · Guest agent not running", with History deferred; after, the agent's bars, nothing deferred. - Silent agent: before, "- Unavailable"; after, bars plus "Last known" (phone: "Prior"), with History deferred on the agent message. - Stopped guest: Unavailable. - A guest with Proxmox filesystems: unchanged. - Machines row for a standalone agent released from a silent hold, watched without a reload: on main it kept "(stale)"; now it clears. - Bundle: WorkloadsSurface 61.93 kB gzip against a 70.23 kB cap. ## Known residuals - Mock mode's websocket state does not apply manual links to resources ingested as records (`ingestRecords` never runs `applyManualLinks`), so the browser pass ran the Proxmox page on REST data. In production this likely affects links to vSphere or TrueNAS resources; it is tracked separately. - A stale Proxmox source can lose the aggregate to a live agent while its own filesystem rows and `prev-` reason remain. That guest stays on the Proxmox rule and reads its live aggregate as last known, which is the same as main and errs toward caution. A guest with no platform read reason can likewise keep a retained aggregate after the agent withdraws its filesystems, also as on main. Both need aggregate provenance on the wire. |
||
|
rcourtman
|
15322e1573 |
Drop contract prose for the deleted problem-resource surfaces (#2706)
Contract prose still described surfaces that no longer exist: - `frontend-primitives.md` told new route-level copy to extend `problemResourcePresentation` (deleted in |
||
|
rcourtman
|
23bf99e1ba |
Keep alert settings a config save leaves out (#2689)
## Problem
Saving anything on Alerts > Thresholds turned flapping detection and
alert TTL cleanup off. The page's `PUT /api/alerts/config` body never
carries `flappingEnabled`, `maxAlertAgeDays`, `maxAcknowledgedAgeDays`,
`autoAcknowledgeAfterHours` or `customRules` (the page has no control
for any of them). `UpdateAlertConfig` decoded the body into a zero
`AlertConfig`, so every save wrote `false`/`0` for those fields, reset
the flapping tuning to defaults, dropped custom rules, and persisted
that. The `flapping`, `aggregation` and `ioNormalization` objects the
page sends are not `AlertConfig` fields and never enabled anything.
## Fix
- `PUT /api/alerts/config` now replaces only the top-level keys the body
carries. A key left out keeps its stored value. A sent key replaces its
setting whole, so `"overrides": {}` still clears overrides, and the
usual normalization still runs. A body that is not a JSON object
(including `null`) is a 400 and changes nothing.
- `alertconfig.ApplyAlertConfigUpdate` decodes the body on its own, so
sent keys get exact `encoding/json` semantics, including duplicates and
case-folded keys. It copies every unsent top-level field from a JSON
clone of the stored config.
- `Manager.ApplyConfigUpdate` reads, merges and applies under the
manager lock, so concurrent partial saves cannot revert each other. It
returns an owned snapshot, which the handler persists, so
`SaveAlertConfig`'s in-place normalization no longer writes into the
live config's maps. The handler runs one save at a time, so successful
saves reach `alerts.json` in the order they were applied.
- Stored configs are not rewritten; a value an earlier save zeroed stays
as it is.
## Verification
- Unit: `internal/alerts/config/update_test.go`; manager save path and
concurrent partial updates in
`internal/alerts/config_validation_test.go`; handler against a real
manager in `internal/api/alerting/alerts_test.go`. Each new regression
fails against main's code (overlay negative controls). The concurrency
test fails at round 0 against a non-atomic variant.
- `go test ./internal/alerts/... ./internal/api/alerting/` and the
`internal/api` alert endpoint, tenant and scope tests pass; the alert
tests pass under `-race`.
- Private stack (mock mode, Vite on main's frontend):
- On main, a node CPU change saved from the Thresholds page took
flapping and all three TTLs to false/0, in memory and in `alerts.json`.
- With this change, the same UI save kept them (true/7/1/24) and stored
the new threshold.
- The API repro (GET, drop the four keys, PUT) keeps them; explicit
`false`/`0` still applies; `null` and `[]` return 400.
- Five rounds of concurrent partial PUTs kept both changes in memory and
on disk.
## Follow-ups
- The thresholds page still sends the dead
`flapping`/`aggregation`/`ioNormalization` objects and hard-coded
`minimumDelta`/`suppressionWindow`/`hysteresisMargin`. A frontend
cleanup will follow once the open PRs touching
`alertsConfigurationModel.ts` land.
- `SaveAlertConfig` still normalizes caller-owned maps in place for
other callers (`ActivateAlerts`, startup paths), and the GET handler
encodes a shallow copy. That pre-existing race is tracked separately.
|