GitHub has attributed its 7-hour-47-minute outage on August 17 to a Central US infrastructure component that failed to scale when traffic reached a new peak. The resulting capacity shortage cascaded into authentication, GitHub Actions, APIs, pull requests, issues, Pages, webhooks, and Copilot, while a client-side retry loop prolonged recovery.
Source