GitHub favicon

GitHub outage on 2026-11-09

github.com
November 2026 9 Major incident

GitHub had four separate incidents on 9 November 2026, touching Webhooks, Actions, GitHub Pages and Copilot at different points in the day. The longest ran 3h 14m and was GitHub's doing, tied to a surge in anonymous traffic hitting its servers.

Started

10:00 UTC

Duration

Lasted 1h 52m

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: Copilot, Webhooks, Actions, GitHub Pages

What happened?

The first and longest incident started at 05:59 UTC and ran for 3h 14m. GitHub's status page describes a spike in anonymous traffic causing 504 Gateway Timeout errors for users who were not logged in, including CI jobs and release downloads making anonymous calls. GitHub applied a mitigation around 06:51, saw the problem return at 07:43, and did not reach a final resolution until 09:09. Throughout, GitHub stated logged-in users and authenticated requests were unaffected. A second incident began at 10:00 UTC and lasted 1h 52m, hitting Webhooks and Actions with delayed runs and delivery delays, mitigated in stages rather than all at once. A third incident at 18:12 UTC ran 49 minutes and spread across Actions, the Checks API, GitHub Pages and Webhooks, with GitHub confirming all three were mitigated together at 18:43. The fourth, starting at 23:16 UTC, lasted 57 minutes and affected Copilot's GPT-5.3 Codex and GPT-5.6 model options, with GitHub attributing it to an upstream model provider issue and recommending users switch models or select Auto. Oddly, one update in that incident read "Copilot is operating normally" even while the incident was still open, a sign of how these status feeds get updated component by component rather than as one coherent line. None of the four incidents has a root cause analysis published yet; GitHub's boilerplate on each promises one "as soon as it is available". No user reports were filed during the window on any of the four incidents.

Learning

UptimeRobot's own probes against GitHub's api-auth and raw-cdn endpoints from the EU and North America held 100% uptime over the trailing 24 hours and stayed above 99.8% for the week, which can look like a contradiction next to four incidents in one day. It isn't: the failures GitHub described were scoped to anonymous, unauthenticated traffic, to Actions run scheduling, to Webhooks delivery, and to specific Copilot model backends, none of which map directly onto a simple authenticated reachability check hitting a CDN edge. A probe that logs in and fetches a file can pass cleanly while CI jobs time out on anonymous calls, while webhook events queue up late, or while one AI model provider behind Copilot degrades. This is the shape to watch for: narrow, component-specific failures that a broad external check won't catch because it's testing a different path through the same service. Your own monitoring tells you your product is reachable and responding; it says nothing about whether GitHub's Actions queue or Webhooks delivery pipeline is running on time behind the scenes. That's the gap UptimeRobot's provider-side watching is meant to close, so when something breaks you know whether it's your code or GitHub's infrastructure without having to guess.

4.7
stars out of 5
284+ reviews on

Start monitoring in 30 seconds.

There's nothing to install. No credit card required. 50 monitors for free.