fix(update): tolerate late proxy readiness

[luvs01]

Summary

• extend the replacement-proxy arrival budget from 15 seconds to 30 seconds while preserving the existing 15-second stability window
• guarantee one identity-aware health probe at the arrival deadline so a proxy that becomes ready on the boundary is not misclassified as timed out
• add deterministic regressions for readiness at 30 seconds and for the reported post-15-second npm service replacement, including the no-redundant-restart evidence path

This stays intentionally narrower than #533/#557: it does not change process recovery, port reclamation, or restart ownership, and both current PR heads still retain the 15-second arrival loop.

Fixes #720

Verification

• bun test tests/update-job.test.ts (34 passed)
• bun x tsc --noEmit
• bun scripts/privacy-scan.ts
• git diff --check
• The initial monolithic Windows run was cut off by a 10-minute local command budget. Current CI documents an approximately 11.8-minute Windows baseline and uses a 20-minute job budget (CI: windows-latest test job cancelled at the 12-minute ceiling, reported as a failing check #717); isolated current-dev slices continued making progress and completed, confirming a harness cutoff rather than a deadlock. The separate 5-second cli-restore-back test budget found during diagnosis is already covered by open security: harden post-merge service and management boundaries #707. Cross-platform CI still requires repository approval for this PR.

Checklist

• Scope stays focused and avoids unrelated cleanup.
• Docs or release notes were updated when needed. (No documentation change is needed for this bounded readiness correction.)
• Security-sensitive changes were reviewed for secrets, auth, and unsafe defaults.

Summary by CodeRabbit

• Bug Fixes

  • Improved restart health checks to recognize services that become healthy right at the timeout deadline.
  • Extended the restart health-check window from 15 to 30 seconds.
  • Adjusted restart confirmation timing to avoid unnecessary restarts when a replacement service becomes healthy shortly after the previous cutoff.

• Tests

  • Added coverage for deadline health checks and successful replacement-service confirmation.

Review readiness checklist

This PR stays in draft until every box below is ticked. Tick all four boxes once the requirements are met:

• All CI tests are green on my local testing.

• I pushed my PR to the latest dev commit.

• I resolved all correct Codex and CodeRabbit findings.

• My PR is ready for review.

[coderabbitai]

📝 Walkthrough

Walkthrough

The restart health timeout increases to 30 seconds, polling now performs a deadline-accurate final probe, and GUI update tests cover both deadline-boundary recovery and delayed replacement proxy acceptance without redundant restarts.

Changes

Restart health handling

Layer / File(s)	Summary
Deadline-aware restart polling src/update/job.ts	The restart health timeout changes from 15 to 30 seconds; polling dynamically bounds sleep intervals and performs a final probe at the deadline. The decision-log comment describes the 30-second arrival plus 15-second stability window.
Delayed replacement proxy coverage tests/update-job.test.ts	Tests verify the final deadline probe and confirm that a replacement proxy becoming healthy after the previous cutoff succeeds without invoking another restart, while expected log entries are recorded.

Estimated code review effort: 2 (Simple) | ~10 minutes

Sequence Diagram(s)

sequenceDiagram
  participant awaitRestartedProxyHealthy
  participant probeProxy
  participant sleepMs
  awaitRestartedProxyHealthy->>probeProxy: Probe proxy health
  probeProxy-->>awaitRestartedProxyHealthy: Return health status
  awaitRestartedProxyHealthy->>sleepMs: Sleep for bounded remaining interval
  sleepMs-->>awaitRestartedProxyHealthy: Reach deadline
  awaitRestartedProxyHealthy->>probeProxy: Run final deadline probe

Loading

Possibly related PRs

• lidge-jun/opencodex#515: Updates the related GUI restart and health-probing path in src/update/job.ts with corresponding test coverage.

Suggested labels: bug

Suggested reviewers: wibias, lidge-jun, ingwannu

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	The code change matches #720 by extending the readiness window, adding a final probe, and covering the false-failure regression.
Out of Scope Changes check	✅ Passed	The diff stays focused on update-job readiness logic and targeted tests, with no obvious unrelated changes.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly matches the main change: the update job now tolerates proxy readiness that arrives late.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands.}

[lidge-jun]

Merged into dev as 62e937614 (head 11069f055).

Verified locally on the merged tree:

• The loop change is the substantive part, not the 15s→30s constant. awaitRestartedProxyHealthy now always makes one final identity-aware probe at the deadline instead of exiting the while (now() < startDeadline) test first, and the sleep is clamped to the remaining time so it cannot overshoot past the boundary. A replacement that becomes healthy on the last tick is no longer misreported as a timeout.
• Falsification: 34 pass / 0 fail on the merged tree. Reverting only src/update/job.ts to the pre-merge cf30878fc gives 32 pass / 2 fail; restoring it returns 34 pass / 0 fail. The 15.25s arrival case in tests/update-job.test.ts is doing real work.
• Full suite on the merged tree: 6080 pass / 0 fail, 6081 tests across 438 files.

The honest limit, which your own doc comment already records: a replacement arriving later than 30s is still reported as a restart failure. This narrows the false-positive window rather than removing it, and the worst case now takes 30s arrival plus a 15s stability window before success is declared. That tradeoff is the right one for an update path where a wrong "restart failed" costs the user more than a slower confirmation.

Closing as merged. Issue #720 is closed with the same evidence.

[github-actions]

⏳ DRAFT

• review readiness checklist open (0/4 boxes ticked).

What to do

• Tick all four boxes in the PR description once you're done (currently 0/4).

Review readiness checklist

• ⬜ All CI tests are green on my local testing.
• ⬜ I pushed my PR to the latest dev commit.
• ⬜ I resolved all correct Codex and CodeRabbit findings.
• ⬜ My PR is ready for review.

0/4 boxes ticked.

Automatic draft conversion failed. Please convert this pull request to a draft manually until every box above is ticked.