GitHub Actions, Issues and PRs went down for five hours
A status-page incident stalled Actions runs and the Checks API for over two hours today, and left web and email notifications backed up for longer still.
GitHub opened an incident at 12:21 UTC today for degraded performance across Actions, Issues and Pull Requests. Users couldn’t reliably load the Actions tab UI, queue new Actions runs, call the Checks API, or create and update issues, and PR review requests failed over the API too.
Pull Requests recovered fastest, reported back to normal by 13:02 UTC. Actions and Issues took longer: a mitigation attempt at 13:38 UTC didn’t hold, and the degradation wasn’t declared fixed until 14:35 and 14:36 UTC, over two hours after the incident opened. Web and email notifications lagged furthest behind, still working through a backlog as of GitHub’s 17:34 UTC update, five hours after it started. GitHub has not published a root cause.
Why it matters: Actions is the CI layer most of the industry builds merge gates and deploys on. Two-plus hours of failed run queueing and a broken Checks API means stalled merges and deploys for anyone with required status checks, and the notification backlog meant some teams would not even have seen a failing check or a review request until well after the fact.
The caveat: no cause has been given yet, so there’s no way to tell from this report alone whether it was a one-off or something worth watching for a repeat.