GitHub suffered a widespread outage on August 17 that disrupted core services including its API, Actions, Pull Requests, Issues, Pages, Webhooks, and Copilot, with some repository downloads experiencing error rates of around 50%.
The incident began at 1:40 p.m. UTC when GitHub said it was investigating reports of performance problems affecting some of its services. The disruption subsequently spread across multiple parts of the platform before being resolved at 9:15 p.m. UTC, roughly eight hours later.
At the height of the incident, GitHub reported error rates of approximately 20% for its web experience and API traffic. Archive downloads and raw repository content downloads were affected more severely, with error rates reaching around 50%. SAML and OIDC authentication, SCIM, and Team Sync were also impacted, potentially preventing some enterprise customers from accessing their development environments.
The disruption affected a broad range of developer workflows. GitHub Actions, which is used to automate software builds, testing, and deployments, experienced degraded availability, while Pull Requests, Issues, Pages, Git Operations, and Webhooks were also affected. GitHub Copilot, the company's AI coding assistant, experienced degraded availability during the outage as well.
GitHub began reporting signs of recovery after identifying the problematic component and taking corrective action at around 4:36 p.m. UTC. However, several services continued experiencing problems, including sporadic authentication failures. The company later said it had partially disabled authentication token retries to mitigate the remaining issues.
The outage generated thousands of user reports, with Downdetector recording nearly 3,000 reports at one point. GitHub's status page ultimately marked the incident as resolved at 9:15 p.m. UTC, although the company said it would provide a detailed root cause analysis once available.
The incident comes as GitHub faces rapidly increasing demand from AI-assisted software development. The company has previously said that repository creation, pull request activity, API usage, automation, and workloads involving large repositories are all growing rapidly as developers increasingly use AI coding tools.
GitHub began working to increase infrastructure capacity by 10 times last year but later determined that it would need a 30-fold increase to accommodate the changing workload, according to CTO Vladimir Fedorov. The company has also been working to improve caching, isolate critical services, remove single points of failure, and move performance-sensitive workloads to infrastructure better suited to the increased demand.
The latest outage also follows several other disruptions affecting GitHub, with the company reporting 26 incidents in July 2026 alone. GitHub has not yet disclosed the specific cause of the latest incident, noting that a detailed analysis will be shared once it becomes available.
Leave a Reply