Repository navigation
ci: cancel superseded PR runs and give every job a timeout - #268
Merged
Merged
Conversation
A PR's CI is about 200 job-minutes (Test windows ~57m, macOS ~45m, ubuntu ~37m median). Nothing cancelled the run for a superseded push, so each push to a PR stacked a full run on top of the one still going. No job set timeout-minutes either, so a hung test held its runner (possibly one of our few macOS slots) for GitHub's 6-hour default. - workflow-level concurrency group per PR with cancel-in-progress - timeout-minutes on every job, set above the slowest successful run in the last ~50: test 90 (max seen 76), build 30 (14), filesystems/no-logging/devnet 25 (9.5), the rest 20, fmt 10 Closes V2-1822. The job consolidation follows separately. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2 of 7 tasks
MaidSafe-Hermes
approved these changes
Nov 3, 2026
MaidSafe-Hermes
left a comment
MaidSafe-Hermes
left a comment
There was a problem hiding this comment.
APPROVE
Reviewed WithAutonomi/ant-node@46ebc21b0f69cf9af81a684b9bc6731c4d0fc515 against base 02c8682466df5c0c6723f32c46223a572f020ed4. The workflow-level per-PR concurrency group and all 11 job timeouts are valid and match the intended behavior. No blocking code issue found.
Coverage: 2/2 seats reviewed (sol-1 gpt-6.1-sol xhigh; glm-1 z-ai/glm-5.3); both seats agree (two model families). No local tests run; this CI-only change was reviewed statically. Current Security Audit and Third-party notices failures are separate from this diff and are merge-check concerns, not code-review blockers.
grumbach
added a commit
to grumbach/ant-node
that referenced
this pull request
Nov 3, 2026
Brings in the browser listener's UPnP port mapping (WithAutonomi#264, ADR-0027), the CI concurrency and job timeouts (WithAutonomi#268) and the removal of claude.yml (WithAutonomi#267). Conflicts: the endpoint publication, where main's Publication replaces the single endpoint and the swarm's STUN server is still set beside it, and the ADR index, which lists both ADR-0026 and ADR-0027. A mapped endpoint withdrawn when its mapping ends now also clears the swarm's STUN server, so the node hosts no rooms while browsers cannot reach it, as before any endpoint is published. ADR-0026 says so, and main's two withdrawal tests check the STUN server is set on publication and gone after withdrawal.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Linear issue
Closes V2-1822
Summary
The first, small step of the ant-node CI cost work (parent V2-1811, from the CI audit V2-1799). It's split out so it can merge ahead of the job consolidation, which follows as a separate PR stacked on this one.
A PR's CI costs about 200 job-minutes (median: Test windows 57m, macOS 45m, ubuntu 37m). Two gaps made that worse:
timeout-minutes. A hung test held its runner, possibly one of the org's few macOS slots, for GitHub's 6-hour default.This PR:
concurrencygroup per PR (ci-<PR number>) withcancel-in-progress: truetimeout-minutesto every ci.yml job, set above the slowest successful run in the last ~50:Not touched:
pr-checks.yml(kept verbatim across six repos),adr-governance.yml,release.yml.Risk tier
Compatibility
Semver impact
Test evidence
timeout-minutes, and theconcurrencyblock is present.New dependency
none
ADR
n/a
Mitigation / rollback
Revert the commit. If a timeout proves too tight on a cold cache, raise that job's value.
🤖 Generated with Claude Code