Repository navigation
ci: GitHub Actions stopped assigning runners; 126 runs queued since 2026-10-22 23:50Z #7572
Description
Client or integration
Other
Area
Installation or packaging
Summary
GitHub Actions has stopped assigning runners to jobs in lidge-jun/opencodex. The last job that started on a runner began at 2026-10-22T23:50:31Z (desktop shell in run 39870500860, which is still in_progress). Since then no other job has started, and 126 runs are queued (GET /repos/lidge-jun/opencodex/actions/runs?status=queued total_count=126 at 2026-10-23T00:22Z). Runs that need no runner (skipped / action_required) still complete immediately. githubstatus.com reports Actions as operational.
This blocks PR CI (for example #7570 and #7571, every check pending) and the Windows signing test for #7207 (run https://github.com/lidge-jun/opencodex/actions/runs/39875468528, dispatched 2026-10-23T00:00:53Z, both jobs queued).
Reproduction
gh run list -R lidge-jun/opencodex --status in_progressreturns 1 run.gh api 'repos/lidge-jun/opencodex/actions/runs?status=queued&per_page=100' --jq .total_countreturns 126.gh run view 39875468528 -R lidge-jun/opencodex --json jobsshows both jobsqueuedwith no runner after more than 20 minutes.
Possible causes to check: an account-level concurrency or spending setting, a stuck job holding the queue, or an unannounced GitHub incident. Cancelling stale queued runs from superseded heads may also help.
Version
dev @ c58e74e
Operating system
GitHub-hosted runners (ubuntu-latest, windows-latest)
Provider and model
n/a
Logs or error output
No response
Screenshots and supporting files
No response
Redacted configuration
No response
Checks
- I searched existing issues and documentation.
- I removed secrets, tokens, account details, request credentials, and personal data.
Activity
lidge-jun commentedon Oct 23, 2026
Update: runner allocation resumed around 2026-10-23T00:32Z without any intervention (the Windows signing test run 39875468528 started then and finished green at 00:41Z). The cause is still unknown, so this stays open in case the stall recurs. Close it if the backlog drains normally.
lidge-jun commentedon Oct 23, 2026
Runners are being assigned again. Since this was filed, hosted runs on the sweep PRs have started and completed normally (for example the full PR checks on #7566 at
752a9bdc2andci.yml lane=all39881826580 on #7567), and the queue is back to ordinary depth (3 in progress, 23 queued at the time of this comment). Part of the backlog was duplicate gate runs and supersededlane=alldispatches, which were cancelled. Closing; reopen if runner assignment stalls again.