GitHub's Pull Requests and Actions degraded for 4h 27m on 16 November 2026, starting at 11:59 UTC. The fault was on GitHub's side, with merge timeouts climbing well above baseline before the company mitigated it.
Started
11:59 UTC
Duration
Lasted 4h 27m
Source
IsDown
Next time
GitHub is degraded. Is your product working?
An incident on their side does not always mean an outage on yours, and the only way to know is to already be watching your own endpoints. Set that up in under a minute and you will be notified the next time your own site is affected. And there will be a next time.
50 monitors for your own websites. No card.
Affected components: Pull Requests, Actions
What happened?
GitHub opened the incident at 11:59 UTC, flagging degraded performance for Pull Requests, and within a minute clarified the specific symptom: elevated timeouts when merging pull requests, with some users unable to merge at all. By 12:35 UTC GitHub put a number on it, 2 to 3% of merges timing out against a baseline near 0.2%. That rate did not improve for hours. At 13:24 and 14:06 UTC GitHub reported mitigations had not restored normal performance, still around 3%. At 15:03 UTC the rate had worsened to 4.75%. At 15:33 UTC GitHub added Actions to the incident, saying it was also experiencing degraded performance. Signs of improvement appeared at 15:46 UTC, with the timeout rate down to 2.5%, and by 15:55 UTC GitHub said it was reducing load on affected systems while recovery was still unconfirmed. GitHub moved to monitoring at 16:07 UTC, stating the degradation to both Actions and Pull Requests had been mitigated, and confirmed recovery at 16:16 UTC. The incident was marked resolved at 16:26 UTC, with GitHub saying a root cause analysis would follow. No such analysis is included in anything shown here, so the cause is not on the record. No user reports were filed during the window on this page, though GitHub's own figures describe a measurable, sustained spike in merge failures.
Learning
UptimeRobot's probes against GitHub's api-auth and raw-cdn components, checked every 60 seconds from both EU and NA regions, show 100% uptime over 24 hours and 7 days, and 99.964% over 30 days, with this incident not registered against those checks. That is not a contradiction. The failure here sat inside merge processing and Actions workflow starts, a layer that simple reachability checks on auth or CDN endpoints do not exercise, so a clean probe history and a real, hours-long degradation in pull request merges can both be true at once. This is the common shape of platform incidents: the edges stay up while a specific internal workflow slows or fails intermittently, visible only through the vendor's own metrics like the 0.2% to 4.75% timeout swing GitHub published. If you were only watching your own endpoints that call GitHub, you would have seen failed merges and slow workflows without any outside signal telling you whether the fault was yours or theirs. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.