fix: identify local worker timeouts in channel replies by gokay-ai · Pull Request #155675 · openclaw/openclaw

gokay-ai · 2026-09-22T10:42:56Z

What Problem This Solves

Fixes: external channel replies blame the selected model provider for HTTP 408 when a typed local worker timeout caused the failure without an HTTP request.

User Impact

User impact: the warning identifies a local worker task timeout instead of an upstream HTTP failure. Genuine provider HTTP 408 warnings retain their attribution. This does not fix the underlying worker timeout or the reported 60-second context-preparation slowdown.

Why This Change Was Made

The existing reply-copy owner now checks the preserved typed worker cause, while original HTTP status facts take precedence. It does not infer local origin from matching message text or reclassify other worker error codes.

This is a user-facing copy correction only: the documented HTTP-like diagnostic status mapping, fallback and retry behavior, raw error causes, and earlier authentication, CLI, storage, and heartbeat handling remain unchanged. It adds no configuration, error metadata, or public API.

Evidence

A normal isolated Gateway, a registered external channel plugin, a real worker-pool timeout, and an independent HTTP recipient reproduced the original warning with zero provider HTTP requests. On candidate 42d6891d16aa59a23c17d9065030879ca5c9efaa, the recipient received one reply: A local worker task timed out. Please try again. The original typed error and diagnostic classification were retained.
An actual loopback HTTP 408 control delivered byte-identical provider-attributed warning text before and after the repair. It recorded eight responses on the baseline and seven on the candidate; retry settings were unchanged, and count equality is not claimed.
The direct and wrapped local-timeout regressions fail on the original copy owner while four controls pass. The candidate passes 137 focused tests covering failure copy, provider handling, and existing priority paths (54.78s wall; the two changed local-origin cases took 1ms each). Core production/test types, scoped lint, line/assertion ratchets, and an exact-source runtime build also passed. Hosted timing for the prepared head is not yet available.
This is generic registered external-channel proof, not Telegram Test Server proof or a reproduction of the native Codex context slowdown. The earlier WebChat capture used its separate internal error copy and did not show an HTTP-labelled warning; that limitation is retained rather than presented as the reproduction.

Fixes openclaw#155627 Signed-off-by: GokayAI <[email protected]>

clawsweeper · 2026-09-22T10:43:04Z

🦞👀
ClawSweeper picked this up.

Pull request received. I will update this pull request when review starts.

ClawSweeper review complete

ClawSweeper finished reviewing this revision. The review result is being finalized.

View the workflow run.

Copilot

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

clawsweeper · 2026-09-22T10:53:54Z

Codex review: needs maintainer review before merge. Reviewed September 23, 2026, 12:10 AM ET / 04:10 UTC (Revision 4).

ClawSweeper review

What this changes

The PR makes channel failure replies identify typed local worker timeouts while preserving provider HTTP error wording, and adds regression coverage.

Merge readiness

✅ Ready for maintainer review

This remains a useful, focused correction absent from current main and the latest release. No blocking correctness or security finding was identified, and the supplied production-path evidence supports the changed reply behavior.

Priority: P2
Reviewed head: 11c037106856743ef35244b54648443a13081002

Review scores

Measure	Result	What it means
Overall readiness	🐚 platinum hermit (4/6)	A focused correction with meaningful regression controls and relevant production-boundary observations, with no blocking finding.
Proof confidence	🐚 platinum hermit (4/6)	Sufficient (live_output): The supplied Gateway run exercised the changed channel reply owner through a registered channel and real worker timeout, recording the corrected recipient text and an unchanged real HTTP 408 control. The changed files match the proof candidate; Telegram-specific delivery and the underlying slowdown are outside that evidence. No stored-data contract changes.
Patch quality	🐚 platinum hermit (4/6)	No actionable review findings were identified.

Verification

Check	Result	Evidence
Real behavior	Verified	Sufficient (live_output): The supplied Gateway run exercised the changed channel reply owner through a registered channel and real worker timeout, recording the corrected recipient text and an unchanged real HTTP 408 control. The changed files match the proof candidate; Telegram-specific delivery and the underlying slowdown are outside that evidence. No stored-data contract changes.
Evidence reviewed	9 items	Pinned introduced change: The merge-base-to-head diff changes only the shared failure-copy owner and its tests. The new branch requires a timeout classification and a WorkerTaskError timeout cause; original status/statusCode fields prevent local reattribution. Earlier authentication, CLI, heartbeat, and curated failure handling retain precedence. Current main still needs the correction: The fetched main revision has the same pre-fix reply owner as the pinned base: classified local timeouts still reach the provider-attributed failure renderer. This comparison establishes current behavior, not ownership of unrelated base changes. Latest release check: The reply owner in v2026.9.5 also proceeds directly from existing special-case handling to classified provider copy, without the local-worker correction. The release blob was read through GitHub after its local object was unavailable.
Findings	None	None.
Security	None	None.

How this fits together

OpenClaw’s shared channel reply formatter turns agent-run errors into user-facing messages. It receives preserved error causes and classified failure facts, then supplies reply text to channel delivery.

flowchart TD
  A[Agent run fails] --> B[Preserved error causes]
  B --> C[Shared channel reply formatter]
  C --> D{Typed local timeout without original HTTP status?}
  D -->|Yes| E[Local worker timeout message]
  D -->|No| F[Existing failure guidance]
  E --> G[Channel delivery]
  F --> G

Before merge

None.

Agent review details

Security

None.

PR surface

Source +30, Tests +49. Total +79 across 2 files.

View PR surface stats

Area	Files	Added	Removed	Net
Source	1	31	1	+30
Tests	1	61	12	+49
Docs	0	0	0	0
Config	0	0	0	0
Generated	0	0	0	0
Other	0	0	0	0
Total	2	92	13	+79

Review metrics

Metric	Value	Why it matters
Production versus test growth	production +30 net lines; tests +49 net lines	Production growth implements typed-cause attribution and HTTP precedence in the existing owner; tests protect the distinction and diagnostic redaction.

Technical review

Best possible solution:

Keep origin-aware wording in the existing reply formatter while retaining diagnostic classification and tracking the underlying context slowdown separately.

Do we have a high-confidence way to reproduce the issue?

Yes: current source preserves a typed worker timeout while synthesizing status 408, then renders provider-attributed channel copy. The supplied Gateway run corroborates that path; this review did not execute a current-main reproduction.

Is this the best way to solve the issue?

Yes: the patch corrects presentation at its existing owner without changing timeout budgets, fallback policy, public APIs, or stored state.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning medium; reviewed against de32d6b7a12e.

Labels

Label changes:

No label changes.

Label justifications:

P2: Incorrect timeout attribution misleads users, but this bounded copy correction does not repair or introduce turn failures.
rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🐚 platinum hermit and patch quality is 🐚 platinum hermit.
status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (live_output): The supplied Gateway run exercised the changed channel reply owner through a registered channel and real worker timeout, recording the corrected recipient text and an unchanged real HTTP 408 control. The changed files match the proof candidate; Telegram-specific delivery and the underlying slowdown are outside that evidence. No stored-data contract changes.
proof: sufficient: Contributor real behavior proof is sufficient. The supplied Gateway run exercised the changed channel reply owner through a registered channel and real worker timeout, recording the corrected recipient text and an unchanged real HTTP 408 control. The changed files match the proof candidate; Telegram-specific delivery and the underlying slowdown are outside that evidence. No stored-data contract changes.
proof: telegram-e2e: This PR needs Telegram Test Server proof with the repository E2E skill. The shared formatter changes Telegram-visible warning text; supplied generic-channel proof does not establish a Telegram Test Server run.

Evidence

What I checked:

Pinned introduced change: The merge-base-to-head diff changes only the shared failure-copy owner and its tests. The new branch requires a timeout classification and a WorkerTaskError timeout cause; original status/statusCode fields prevent local reattribution. Earlier authentication, CLI, heartbeat, and curated failure handling retain precedence. (src/auto-reply/reply/agent-runner-failure-reply.ts:257, 11c037106856)
Current main still needs the correction: The fetched main revision has the same pre-fix reply owner as the pinned base: classified local timeouts still reach the provider-attributed failure renderer. This comparison establishes current behavior, not ownership of unrelated base changes. (src/auto-reply/reply/agent-runner-failure-reply.ts:379, de32d6b7a12e)
Latest release check: The reply owner in v2026.9.5 also proceeds directly from existing special-case handling to classified provider copy, without the local-worker correction. The release blob was read through GitHub after its local object was unavailable. (src/auto-reply/reply/agent-runner-failure-reply.ts:394, ec9c1a13db89)
Production boundary and dependency scope: The worker pool creates WorkerTaskError with code timeout; failover coercion preserves its Error cause while assigning HTTP-like status 408. The channel error handler passes that preserved error to the changed formatter. This correction depends on OpenClaw’s local error contracts, not the Codex runtime or protocol; the PR explicitly excludes fixing native context preparation. (src/auto-reply/reply/agent-runner-error-handler.ts:259, 11c037106856)
Supplied real behavior evidence: The captured PR body reports an isolated normal Gateway, registered external channel, real worker-pool timeout, and independent HTTP recipient. Candidate 42d6891 delivered one local-worker message with zero provider HTTP requests. A loopback HTTP 408 control retained byte-identical provider wording. This is observed channel-boundary output, not Telegram Test Server evidence or proof of resolving the underlying slowdown. (42d6891d16aa)
Proof and review continuity: Both changed files are identical between the proof candidate and the reviewed head; comparison with the immediately preceding reviewed head is also empty. The previous completed review retained no findings. No historical blocker was silently dropped or inferred from omitted self-comments. (src/auto-reply/reply/agent-runner-failure-reply.ts, 11c037106856)

Likely related people:

obviyus: Suggested for follow-up; no historical authorship or introduction is verified. (role: unverified routing candidate; confidence: low)
steipete: Suggested for follow-up; no historical authorship or introduction is verified. (role: unverified routing candidate; confidence: low)

Rating scale

Score	Internal tier	Crab rank	Meaning
6/6	S	🦀 challenger crab	Exceptional readiness
5/6	A	🦞 diamond lobster	Very strong readiness
4/6	B	🐚 platinum hermit	Good normal PR; ordinary maintainer review
3/6	C	🦐 gold shrimp	Useful, but confidence is limited
2/6	D	🦪 silver shellfish	Proof or implementation needs work
1/6	F	🧂 unranked krab	Not merge-ready
N/A	NA	🌊 off-meta tidepool	Rating does not apply

Overall follows the weaker of proof and patch quality.
Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

Workflow

ClawSweeper keeps one durable marker-backed review comment per issue or PR.
Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
Maintainers can also comment @clawsweeper review to request a fresh review only.
Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

History

Review history (3 earlier review cycles)

reviewed 2026-09-22T10:52:50.902Z sha 18babe8 :: needs real behavior proof before merge. :: none
reviewed 2026-09-22T15:18:53.512Z sha 42d6891 :: needs changes before merge. :: none
reviewed 2026-09-23T03:11:06.334Z sha 0feab48 :: needs maintainer review before merge. :: none

github-actions · 2026-09-22T15:11:26Z

Security-sensitive guard cleared

Current revision: 11c037106856743ef35244b54648443a13081002

This PR no longer changes files in the maintainer security-review tier.

Separate CODEOWNERS requirements still apply to security policy and enforcement files.

obviyus · 2026-09-22T15:29:40Z

/allow-security-sensitive-change

…al-diagnostic

obviyus · 2026-09-23T04:29:02Z

Landed in d7b8c98. Typed local worker timeouts now get local-timeout guidance in external channel replies, while genuine HTTP 408 errors retain provider attribution. The refreshed head passed all 21 focused failure-reply tests, required CI, and hosted review. Retained isolated Gateway/channel proof covers real worker timeout delivery and the HTTP 408 control; this does not fix the underlying slowdown. Thanks @gokay-ai for the contribution!

fix(failover): stop labelling local context timeouts as HTTP 408

18babe8

Fixes openclaw#155627 Signed-off-by: GokayAI <[email protected]>

openclaw-barnacle Bot added agents Agent runtime and tooling size: M labels Sep 22, 2026

gokay-ai marked this pull request as ready for review September 22, 2026 10:45

Copilot AI lite review requested due to automatic review settings September 22, 2026 10:45

Copilot AI reviewed Sep 22, 2026

obviyus added 2 commits September 22, 2026 20:25

fix: distinguish local worker timeouts in channel failure replies

13d5182

Merge original timeout repair while preserving scoped user-copy fix

42d6891

obviyus self-assigned this Sep 22, 2026

obviyus requested a review from a team as a code owner September 22, 2026 15:10

obviyus changed the title ~~fix(failover): stop labelling local context timeouts as HTTP 408~~ fix: identify local worker timeouts in channel replies Sep 22, 2026

github-actions Bot added security-sensitive-changed security-review-required labels Sep 22, 2026

github-actions Bot removed security-sensitive-changed security-review-required labels Sep 22, 2026

Merge remote-tracking branch 'origin/main' into fix/timeout155675-loc…

0feab48

…al-diagnostic

Merge remote-tracking branch 'origin/main' into fix/timeout155675-loc…

11c0371

…al-diagnostic

YangManBOBO mentioned this pull request Sep 23, 2026

fix: attribute local worker deadlines locally while preserving configured fallback #156650

Merged

roboclaw-bot mentioned this pull request Sep 24, 2026

2026.9.7 Fixes Tracker #157531

Open

2 of 8 tasks

clawsweeper Bot mentioned this pull request Sep 26, 2026

[Bug]: Local Codex context-worker timeout aborts Telegram turns and surfaces as model HTTP 408 #155627

Open

Navigation Menu

Repository navigation

fix: identify local worker timeouts in channel replies - #155675

fix: identify local worker timeouts in channel replies#155675
obviyus merged 5 commits into
openclaw:mainfrom
gokay-ai:cursor/local-context-timeout-http-408-9f7a

gokay-ai commented Sep 22, 2026 •

edited by obviyus

Loading

Uh oh!

clawsweeper Bot commented Sep 22, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

clawsweeper Bot commented Sep 22, 2026 •

edited

Loading

Security

PR surface

Review metrics

Technical review

Labels

Evidence

Rating scale

Workflow

History

Uh oh!

github-actions Bot commented Sep 22, 2026 •

edited

Loading

Uh oh!

obviyus commented Sep 22, 2026

Uh oh!

obviyus commented Sep 23, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

gokay-ai commented Sep 22, 2026 • edited by obviyus Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What Problem This Solves

User Impact

Why This Change Was Made

Evidence

Uh oh!

clawsweeper Bot commented Sep 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

ClawSweeper review complete

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Uh oh!

clawsweeper Bot commented Sep 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

ClawSweeper review

What this changes

Merge readiness

Review scores

Verification

How this fits together

Before merge

Security

PR surface

Review metrics

Technical review

Labels

Evidence

Rating scale

Workflow

History

Uh oh!

github-actions Bot commented Sep 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Security-sensitive guard cleared

Uh oh!

obviyus commented Sep 22, 2026

Uh oh!

obviyus commented Sep 23, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

gokay-ai commented Sep 22, 2026 •

edited by obviyus

Loading

clawsweeper Bot commented Sep 22, 2026 •

edited

Loading

clawsweeper Bot commented Sep 22, 2026 •

edited

Loading

github-actions Bot commented Sep 22, 2026 •

edited

Loading