Incident with GraphQL API Requests
GitHub · Last saved provider status: resolved
The last saved event status is not the current service status. Source lists can be incomplete, and disappearance from a source does not confirm recovery. Open the current report or the official source for newer evidence.
Saved event details
- Provider status
- resolved
- Provider impact
- minor
- Provider record created
- 27 Jul 2026, 03:53:19 UTC
- Provider record updated
- 11 Aug 2026, 16:17:32 UTC
- Explicit provider start
- 27 Jul 2026, 03:53:19 UTC
- Explicit provider end
- 27 Jul 2026, 04:09:10 UTC
Affected components reported by the provider
- API Requests
brv1bkgrwx7q
These associations describe this event’s reported scope. They do not establish current component availability or verified dependencies.
A record creation time is not necessarily an outage start. Unreported times remain unavailable. We do not calculate downtime from collection times.
Provider updates in saved revisions
Latest first. We show up to 100 distinct updates from the latest 20 saved content revisions. Revised wording at the same provider time is retained separately.
- resolved
On July 26, 2026 at 21:34 UTC we began seeing intermittent errors on the GitHub GraphQL API. A subset of GraphQL API requests returned HTTP 502 errors in short bursts. During the impact window an average of 0.09% of GraphQL API requests in the affected region failed, with a peak of 0.50% of requests failing during the worst two-minute period at 03:02 UTC on July 27. Requests that failed generally succeeded when retried, and no data was lost or altered. Other GitHub services were not affected.<br /><br />The errors were traced to a single group of servers handling a share of GraphQL API traffic. Application processes on that group intermittently closed connections before completing responses. Impact ended at 03:52 UTC on July 27 when those processes were replaced, and we resolved the incident at 04:09 UTC on July 27 after confirming error rates had returned to normal.<br /><br />We are still investigating why those processes closed connections, and that work is being carried out by the team that owns the underlying compute platform. In the meantime we are adding detection and automated mitigation for when a single group of servers behaves differently from its peers.
Seen in a saved revision at 06 Oct 2026, 13:58:39 UTC - monitoring
The degradation affecting API Requests has been mitigated. We are monitoring to ensure stability.
Seen in a saved revision at 06 Oct 2026, 13:58:39 UTC - investigating
We are investigating reports of degraded performance for API Requests
Seen in a saved revision at 06 Oct 2026, 13:58:39 UTC