The GitHub outage on August 17 affected GitHub.com services including Issues, Pull Requests, APIs, Actions and Copilot, according to the company’s status record. GitHub said the incident ran from 13:28 to 21:15 UTC, a total of seven hours and 47 minutes, before the incident was resolved.

GitHub’s incident report said peak error rates reached about 20% for web and API traffic and about 50% for archive and raw-content traffic. SAML and OIDC authentication, SCIM and Team Sync were also affected. Most services recovered by 16:36 UTC, while Actions remained degraded until about 18:03 UTC and Copilot Token Service recovered fully at 21:02 UTC.

The company said it moved some traffic from its Central United States region to Northern Virginia during the incident. GitHub also reported that delayed replies exposed a retry issue in Visual Studio Code, which amplified traffic to the Copilot Token Service by about ten times.

Windows Central reported that GitHub and Copilot users were seeing failures and delays during the incident, while GitHub’s own status page recorded the service-by-service recovery. The status page currently lists all systems as operational. The available reports describe a resolved service incident and do not indicate that the August 17 disruption is still active.