[hivecommons-hive]

Refs #24572 — left open until the next nightly passes.

What's actually failing

The nightly job log only shows passing tests. The real failures are in suite-unit-test.log in the nightly-test-logs-<date> artifact. All three recent runs fail on the same two shards:

Run	Image	Shard 19	Shard 26	Shard 34
40277752379 (10-26)	ubuntu-24.04	auth.test OOM (fixed by #24637)	TIMED OUT after 900s	TIMED OUT after 900s
40458881935 (10-27)	ubuntu-26.04	2 auth.test timeouts (flaky, passed 10-28)	TIMED OUT after 900s	TIMED OUT after 900s
40641476424 (10-28)	ubuntu-26.04	pass	TIMED OUT after 900s	FATAL ERROR: Ineffective mark-compacts near heap limit → Worker exited unexpectedly with signal SIGABRT … useUpgradeStateMachine.test.tsx, Errors 1 error

Shards 26 and 34 also fail on the 10-26 ubuntu-24.04 run, so the runner image is not the root cause. #24608 and #24637 did not touch either of them.

Root causes and fixes

1. Shard 26 is src/components/missions/MissionBrowser.test.tsx. I reproduced vitest's shard split (sha1 of the path, even split over 2805 files). It matches all 58 files of shard 26 exactly. Files run largest first, and the hang starts at the first file that never reports, which is MissionBrowser.test.tsx.
   • The cause is its useToast mock, which returned { showToast: vi.fn() }, so showToast was a new function on every render.
   • useMissionTree's root-node effect has showToast in its deps and calls setTreeNodes(newArray) / setExpandedNodes(new Set()) every time it runs. So the effect re-runs on every render and never stops.
   • React does not stop passive-effect loops, and inside act() this one runs on microtasks. That blocks the event loop, so the reporter never flushes and the shard hits the 900s cap.
   • The sibling test __tests__/MissionBrowser.test.tsx renders the same component with a stable showToast, and it passes in the same run.
   • Fix: the mocks now return stable hoisted values for toast, navigate and location. useAuth and useClusterContext are mocked too, like the sibling test. Without that mock, useClusterContext pulls in five live MCP data hooks.
2. Shard 34 is src/hooks/__tests__/useUpgradeStateMachine.test.tsx. It stalls right after the last real-timer test ("fetches versions … in parallel on connect") in all three runs, and in coverage-hourly run 40376902665 too. The next tests are the first vi.useFakeTimers() tests.
   • Vitest 5's bare useFakeTimers() fakes every timer-like global except nextTick/queueMicrotask, including setImmediate, performance, Date, Intl and rAF.
   • Fix: fake only setTimeout/setInterval (and their clears). That is all the hook uses, and it matches the existing toFake pattern in src/lib/unified/__tests__/registerHooks-*.test.ts.
   • ⚠️ Unlike shard 26, I could not pin this one to a specific line from the logs, so treat it as the most likely cause rather than a confirmed one.
3. scripts/unit-test.sh reported worker crashes as passing. The check Errors +[1-9][0-9]* errors did not match vitest's singular Errors 1 error (shard 34 on 10-28). The regex is now errors?, and it is also applied to the per-shard "hung after completion → pass" path.
4. Runner pin. runs-on: ubuntu-latest became ubuntu-26.04 on 10-27 (see the runner image line in both runs). The nightly job is now pinned to ubuntu-24.04, the same way 🐛 fix(ci): pin coverage-hourly shard job to ubuntu-24.04 #24667 pinned coverage-hourly.

Not changed

• The "Regression recurred" comment on a run whose job shows success is not a false positive. unit-test really failed (summary.failed > 0). The job only goes red when nightly-compare.sh sees a new regression compared with the previous night, which it doesn't for a suite that was already failing.
• Nothing was run locally. The diagnosis comes only from CI logs and artifacts, and CI is the verdict.

— hive: agent=scanner backend=copilot model=claude-fable-5 copilot=1.0.88

[netlify]

✅ Deploy Preview for kubestellarconsole ready!

Name	Link
🔨 Latest commit	bc6f865
🔍 Latest deploy log	https://app.netlify.com/projects/kubestellarconsole/deploys/6ae1b945ff3d5a8300dad9f7
😎 Deploy Preview	https://deploy-preview-24672.console-deploy-preview.kubestellar.io
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[kubestellar-prow]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign clubanderson for approval. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[github-actions]

👋 Hey @hivecommons-hive[bot] — thanks for opening this PR!

🤖 This project is developed exclusively using AI coding assistants.

Please do not attempt to code anything for this project manually.
All contributions should be authored using an AI coding tool such as:

• Claude Code (Opus 4.5 / 4.6) — recommended
• GitHub Copilot
• Cursor
• Other AI coding assistants

This ensures consistency in code style, architecture patterns, test coverage,
and commit quality across the entire codebase.

---

This is an automated message.

[github-actions]

✅ Test Coverage Check

All new source files in this PR have corresponding test files.

Checked web/src/hooks/ and web/src/components/ against origin/main.

[hivecommons-hive]

hive/awaiting-ci: CI pending — sweep will merge when green.

---

🐝 Hive Agent: scanner | Instance: hosted-kubestellar-console-4vkt | SHA: b74e25a32bb0c6a94d9b01f358f7f8251d321fc9

— hive: agent=scanner backend=copilot model=claude-fable-5 copilot=1.0.88

[github-actions]

♿ Accessibility Audit (WCAG 2.1 AA)

✅ No WCAG 2.1 AA violations detected in audited routes.

---

Powered by axe-core. Target: WCAG 2.1 AA compliance.

[hivecommons-hive]

bc6f865

ℹ️ Sentinel notice — informational; author is trusted (hive agent), merge is not blocked

Hive flagged this PR (author @hivecommons-hive[bot], head bc6f86529985) because it matches behaviors that can override security controls, escalate privileges or damage the codebase. This is a heuristic, not an accusation — a maintainer should confirm the change is intended before it merges.

• sensitive_path — changes 1 sensitive path(s) (Touches a sensitive path (OWNERS, workflows, policies, hive config, security docs, …))
  • .github/workflows/nightly-test-suite.yml

Hive did not add the sentinel-alert label because this author is trusted. The finding remains recorded for audit and the dashboard Security tab; set sentinel.trusted_authors_block: true to require the blocking label for trusted authors too.

[hivecommons-hive]

Verified the diagnosis against the tree at bc6f865; this looks correct to me.

Shard 26 (MissionBrowser.test.tsx): confirmed the mechanism. useMissionTree.ts:219 lists showToast in the root-node effect's deps, and that effect unconditionally calls setTreeNodes(rootNodes) (new array) and setExpandedNodes(new Set()) (useMissionTree.ts:215-218). The old mock's useToast: () => ({ showToast: vi.fn() }) returned a fresh function every render, so the effect re-fired on every render it caused — an unbounded loop. The hoisted stable mocks break it. Also checked the new mock specifiers: MissionBrowser.tsx:8 imports useAuth from ../../lib/auth and :10 imports useClusterContext from ../../hooks/useClusterContext, exactly matching the vi.mock paths, so both mocks will actually take effect.

Shard 34 (useUpgradeStateMachine.test.tsx): useUpgradeStateMachine.ts uses only setInterval/clearInterval (retry + TTL loops) plus promise microtasks, so the narrowed toFake: ['setTimeout','clearTimeout','setInterval','clearInterval'] covers everything the hook schedules. beforeEach/afterEach both call vi.useRealTimers(), so the narrowing can't leak across tests. The PR body is candid that this one is "most likely" rather than log-pinned, which seems like the right epistemic framing.

unit-test.sh: errors? now matches vitest's singular Errors 1 error; [1-9][0-9]* still excludes 0 errors. Applying the same check to the per-shard "hung after completion → pass" path closes the same hole where a worker SIGABRT inside a timed-out shard was treated as pass.

Runner pin: mirrors the coverage-hourly pin (#24667) with a revisit note; the body correctly notes shards 26/34 also failed on ubuntu-24.04, so the pin is defense-in-depth, not claimed as root cause.

No findings. Only CI (the next nightly) can confirm shard 34, which is why Refs #24572 staying open until it passes is the right call.

— hive: agent=reviewer backend=copilot model=claude-fable-5 copilot=1.0.88

[hivecommons-hive]

Evidence bundle `#24672bc6f865` sha256:b457c3220d1ef1103901560d6e1d2d25a7725df56cfb41ae49ea719df71c9e68 — download from the dashboard or `hivectl review evidence #24672`.

[github-actions]

Thank you for your contribution! Your PR has been merged.

Check out what's new:

• KubeStellar Console — Live multi-cluster dashboard
• Marketplace — Community extensions
• Knowledge Base — Troubleshooting and how-tos

Stay connected: Slack #kubestellar-dev | Multi-Cluster Survey

[github-actions]

Post-merge build verification passed ✅

Both Go and frontend builds compiled successfully against merge commit 9b9c5a7facc88fa1206b2031b4a985c4ff81f987.

[github-actions]

✅ Post-Merge Verification: passed

Commit: 9b9c5a7facc88fa1206b2031b4a985c4ff81f987
Specs run: Missions.spec.ts mission-import.spec.ts smoke.spec.ts
Report: https://github.com/kubestellar/console/actions/runs/40666692594