started · updated
GitHub suffers major global outage affecting core developer services
GitHub, the Microsoft-owned code-hosting platform, experienced a major global service disruption on Monday, August 17, 2026. The outage lasted approximately eight hours, significantly impacting software development workflows worldwide.
At the peak of the incident, GitHub reported error rates of roughly 20% for web interfaces and API traffic. The disruption was even more severe for repository downloads, with archive and raw content requests seeing failure rates near 50%. Key services affected included Pull Requests, Issues, GitHub Actions, Webhooks, and the AI-powered coding assistant, GitHub Copilot. Enterprise authentication services such as SAML, OIDC, and SCIM also faced difficulties.
Engineers identified a problematic component within the infrastructure and implemented corrective actions. To assist in recovery, the company partially disabled authentication-token retries to prevent a feedback loop of increased traffic. The incident highlights the immense pressure on GitHub's infrastructure as it scales to meet the demands of AI-driven development, with leadership noting a need for capacity increases up to 30 times current levels.
Entities
Azure · Cursor · GitHub · Microsoft · Origin · SpaceX · Vladimir Fedorov
Claims
What the coverage asserts, and how well corroborated each claim is across sources.
- [● 12 SOURCES] GitHub experienced error rates of approximately 20% for web and API traffic and 50% for archive and raw content downloads. www.brasilemfolhas.com.br · es.digitaltrends.com · www.blog-nouvelles-technologies.fr · borncity.com · crn.pl · +7 more
- [● 3 SOURCES] The outage was caused by a problematic component within the infrastructure. borncity.com · www.computerworld.com · www.windowscentral.com
- [● 12 SOURCES] The outage affected services including Pull Requests, Issues, Actions, Webhooks, and GitHub Copilot. www.brasilemfolhas.com.br · es.digitaltrends.com · www.blog-nouvelles-technologies.fr · borncity.com · crn.pl · +7 more
- [● 2 SOURCES] Authentication-token retries were partially disabled to mitigate the issue. dev.to · www.computerworld.com
- [● 3 SOURCES] The platform is scaling to meet a 30-fold increase in capacity required by AI development. www.brasilemfolhas.com.br · crn.pl · thenewstack.io
- [● 3 SOURCES] The service disruption lasted approximately eight hours. crn.pl · www.computerworld.com · dev.to