fix(#425): scope claude-review.yml's concurrency group per-PR, not per-ref - #426
Conversation
…r-ref github.ref resolves to the default branch for issue_comment and pull_request_review_comment events regardless of which PR the comment is on, so every comment-triggered review across the whole repo shared one concurrency-group queue slot. A newer run entering the group cancels whatever was previously queued (not the in-progress one), so a review requested on one PR could silently evict a still-queued review on an unrelated PR, with no visible error to whoever asked for it. Reproduced live on 2026-08-28: near-simultaneous @claude review comments on PR #422 and #423 collided; #422's queued run was silently cancelled by claude[bot]'s own auto-reply comments on #423 firing into the same shared group. Scope the group by github.event.issue.number (issue_comment, including PR comments) or github.event.pull_request.number (pull_request and pull_request_review_comment), falling back to github.ref for any other trigger. cancel-in-progress is unaffected - it only reads github.event_name. Closes #425 Assisted-by: claude-opus-5 (agent)
|
Warning Review limit reachedNext included review available in 59 minutes. View limit detailsLimit details: You’ve used the included review currently available. You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Review configuration: ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (1)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
There was a problem hiding this comment.
Pull request overview
This PR fixes a GitHub Actions concurrency misconfiguration in the Claude review workflow so that comment-triggered reviews don’t all share a single repo-wide queue slot and cancel each other across unrelated PRs.
Changes:
- Scope
claude-review.ymlconcurrency group by PR/issue number (github.event.issue.number/github.event.pull_request.number) instead ofgithub.ref. - Add explanatory documentation in the workflow clarifying why
github.refis incorrect for comment events and how the new key avoids cross-PR eviction.
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
…ment Reflow the concurrency comment so the code span isn't split across a line break, per Copilot review feedback. Assisted-by: claude-opus-5 (agent)
|
`@claude review |
|
|
Summary
claude-review.yml's concurrency group usedgithub.ref, which resolves to the default branch (not the PR) forissue_comment/pull_request_review_commentevents — every comment-triggered review across the repo shared one queue slot, so a review on one PR could silently evict a still-queued review on a different PR.@claude reviewon PR impl(#406): SSO 4a/4: Login SSO status - data layer (split from #355) #422 and impl(#409): SSO 4d/4: SSO config page + route (split from #355) #423 within ~26s of each other; impl(#406): SSO 4a/4: Login SSO status - data layer (split from #355) #422's queued run was silently cancelled.github.event.issue.number/github.event.pull_request.number, falling back togithub.ref.Test plan
actionlintclean on the changed fileTest WorkflowsCI job (YAML syntax validation) passesCloses #425