Major GitHub Outage: Errors Hit API, Actions, Issues and Copilot

On August 17, 2026, GitHub reported a large-scale outage that affected several of the platform's key services at once. The first notice appeared on its status page at 13:40 UTC: the company had begun investigating reports of degraded performance in a number of services.

Over the next hour, the list of affected components grew quickly. Degraded performance was reported in turn for the API, GitHub Actions, webhooks, and the Issues and Pull Requests sections. By 14:24 the picture had become clearer: the platform was seeing an error rate of roughly 20% across the web interface and API traffic. Downloads had it even worse — fetching archives and the “raw” contents of repositories returned an error in about half of all cases.

What exactly was broken

As the investigation progressed, it turned out that authentication had been affected as well. GitHub listed SAML and OIDC among the impacted mechanisms, along with SCIM and Team Sync — meaning corporate users could face not only a sluggish interface but also problems signing in and syncing teams. The Copilot assistant was caught up in it too, with reduced availability reported for it as well.

By 14:58 UTC the status page noted that the company was already applying mitigations based on the first findings of its investigation and was monitoring how the situation developed. Even so, webhooks, Actions, Pull Requests and Issues were still listed as being in a state of reduced availability, and the cause of the outage was never officially named — the wording amounted to the investigation being ongoing.

Context

Formally, the incident affected six components: webhooks, API requests, Issues, Pull Requests, Actions and Copilot. For a platform that a substantial share of the world's development rests on, even a 20% error rate means noticeable disruption: automated builds and deployments fail to go through, webhooks misfire, and routine operations on code start failing intermittently.

The news drew considerable attention in the community — on Hacker News the discussion quickly gathered hundreds of comments. The reason for this sustained interest is clear: GitHub outages have long served as a kind of barometer of just how heavily the industry depends on a single point of failure, and each such incident is a reminder of the price of that centralization.

Source: Hacker News