fix(codex): a slow but healthy Codex no longer shows as Unavailable

[shivamhwp]

Problem

On a slow machine, the Codex card in Settings > Providers flips to "Unavailable · Timed out while checking Codex app-server provider status" even though Codex is installed, signed in, and working. It stays that way until the next 5-minute refresh. #7513 measured this on Windows in about 1 of 4 probes: the Codex check finished in 10.07s, just past T3's 10s cutoff.

The Codex check does more than a version or auth call. It starts codex app-server, initializes it, and reads the account, models, and skills, all inside the 10s budget every provider shares.

Change

Codex gets its own 20s budget (CODEX_AUTH_PROBE_TIMEOUT_MS) instead of the shared AUTH_PROBE_TIMEOUT_MS. Grok, the only other user of the shared constant, keeps 10s. Nothing wraps checkCodexProviderStatus in a shorter timeout, so the new budget applies end to end.

This rebuilds #9303 by @Michel-Liao on current main. That PR was closed when orchestration V2 replaced the code around it.

Scope and approval

A small fix for an obvious bug: one timeout constant, scoped to the one provider whose check needs it. No behavior changes for working fast machines. Closes #7513.

Verification

Reproduced in the real web app with a wrapper around the real Codex binary that waits 12s before starting (sleep 12; exec codex "$@"), set as the binary path of a Codex instance. The video starts with Refresh provider status and counts the seconds.

Before (main 1e2ecbd): after about 10s the card turns to Unavailable, "Timed out while checking Codex app-server provider status".

9303-before.mp4

After (this PR): the card takes about 14s and lands on Available, v0.160.0.

9303-after.mp4

Focused test: vp test run apps/server/src/provider/ProviderRegistry.test.ts

• New test "does not time out a healthy probe slower than the shared 10s auth-probe timeout ([Bug]: Windows: Codex provider probe exceeds the 10 s AUTH_PROBE_TIMEOUT on ~1 in 4 runs and flips the card to error until the next refresh #7513)" uses a probe that answers after 15s.
• Without the fix it fails: expected 'error' to equal 'ready' (1 failed, 56 passed).
• With the fix: 57 passed (58 after rebasing onto current main). The existing timeout test now advances the clock 21s instead of 11s so it still covers a real timeout.

Not checked: Windows itself. The wrapper adds a fixed delay instead.

Made with Claude Opus 5.5 in Claude Code (T3 Code).

🤖 Generated with Claude Code

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — The PR doubles the hard-coded Codex provider-status timeout from 10 seconds to 20 seconds, changing when a production provider is marked unavailable. The added regression test covers the slower healthy path, while the test-only changes have no runtime impact.

No code changes detected at 9466c36. Prior analysis still applies.

^{You can add or adjust custom eligibility rules. Learn more.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Codex	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Codex	Live turn WebSocket wire	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	20.9 KiB	20.9 KiB	0 B (0.0%)	29.3 KiB	✅
Codex	Live turn messages	2	2	0 (0.0%)	8	✅
Claude	Total thread wire	5.0 KiB	5.0 KiB	+24 B (+0.5%)	6.8 KiB	✅
Claude	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Claude	Live turn WebSocket wire	1.2 KiB	1.2 KiB	+24 B (+2.0%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.2 KiB	21.2 KiB	+41 B (+0.2%)	29.3 KiB	✅
Claude	Live turn messages	1	2	+1 (+100.0%)	8	✅

Baseline: a4c9494 · PR result: 9466c36 · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 108.5 KiB
• Claude decoded thread snapshot: 108.8 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[coderabbitai]

Important

Review skipped

Review was skipped as selected files did not have any reviewable changes.

⚙️ Run configuration

• Configuration used: Path: .coderabbit.config.ts
• Review profile: CHILL
• Plan: Advanced
• Run ID: f423f7e5-2177-4775-a128-fff609ece906

📥 Commits

Reviewing files that changed from the base of the PR and between ccb5cc4 and 9466c36.

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: Path: .coderabbit.config.ts
• Review profile: CHILL
• Plan: Advanced
• Run ID: 91faf920-e0c3-4377-853d-5212c6a14220

📥 Commits

Reviewing files that changed from the base of the PR and between 144f6c8 and ccb5cc4.

📒 Files selected for processing (2)

• apps/server/src/provider/CodexProvider.ts
• apps/server/src/provider/ProviderRegistry.test.ts

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 7 remain after this review.

---

📝 Walkthrough

Walkthrough

The Codex provider status probe now has a 20-second timeout instead of the shared 10-second timeout. Tests cover a healthy probe that takes 15 seconds and a probe that exceeds the new timeout.

Changes

Codex probe timeout

Layer / File(s)	Summary
Codex status probe timeout apps/server/src/provider/CodexProvider.ts, apps/server/src/provider/ProviderRegistry.test.ts	The Codex status check uses a 20-second timeout. Tests verify that a healthy 15-second probe returns ready and that a 21-second probe returns a timeout error. The timeout test also checks that the app-server process is killed once.

Priority: ➖ Normal

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

Change: Bug fix · Severity of issue fixed: Medium

Suggested reviewers: juliusmarminge

Merge Risk: ⚪ Minimal · up to ccb5c

Healthy Codex probes that take between 10 and 20 seconds can now complete instead of showing an error; slower probes still time out. No concrete merge-blocking risk is evident.

🚥 Pre-merge checks | ✅ 4

✅ Passed checks (4 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	Issue #7513 requires the Codex probe to avoid reporting a healthy provider as timed out under normal slow or loaded conditions. The change gives checkCodexProviderStatus a dedicated 20-second timeou…
Out of Scope Changes check	✅ Passed	The supplied change summary contains only the Codex timeout change and focused tests for issue #7513. The test updates verify the new timeout and preserve timeout cleanup coverage. No unrelated change…
Title check	✅ Passed	The title clearly and concisely describes the main change: preventing slow but healthy Codex providers from being marked unavailable.
Description check	✅ Passed	The description covers the problem, change, scope, approval rationale, and focused verification. It includes reproduction details, test results, video evidence, and the limitation that Windows was not…

✨ Finishing Touches

📝 Generate docstrings

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

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

Dismissing prior approval to re-evaluate ccb5cc4