Error rates reach 20%, archive downloads fail 50% while authentication and enterprise integrations experience problems

On the morning of Monday, 17 August 17, 2026 (US time), GitHub, a major platform for hosting and collaborating on code, was reported by numerous news sites to have suffer a major global disruption.

The outage prevented developers in many countries from reaching repositories, merging pull requests, and using automated development tools. Error rates climbed across several of the firm’s most important services, interrupting work for individual programmers and large engineering organizations alike.

The problems began at about 9:40am EDT, when GitHub acknowledged that it was reviewing reports of reduced performance. Over the next hour, the issue expanded beyond the main website to services that developers use throughout the workday. Services such as API Requests, Actions, Webhooks, Issues, and Pull Requests were all affected. GitHub Copilot, the firm’s AI-powered coding assistant, was also experiencing reduced availability by 10:31am EDT.

In an update posted on X, GitHub said web-based experiences and API traffic were seeing error rates of roughly 20%. It had also reported that downloads of archives and raw repository files were failing at an estimated rate of 50%.

The disruption reached enterprise identity-management functions as well: SAML and OIDC authentication, SCIM provisioning, and Team Sync all encountered problems. For business customers that depend on single sign-on, those failures created the possibility that staff would be unable to authenticate and reach their organization’s GitHub resources.

Global outage scenarios

The scale of the incident was visible in reports collected by Downdetector. In the United States, reports peaked at approximately 2,100, while India recorded a high point of 849. About 79% of the Indian complaints concerned access to the GitHub website; other users cited not being able to log in or use the app. Microsoft, GitHub’s parent company, also publicly confirmed that the platform was facing issues around the world. Also:

  • Many software teams use GitHub Actions to automatically build applications, run tests, and deploy releases. During the outage, those continuous integration and continuous deployment processes stalled or returned errors.
  • Automation linked to webhooks had also become unreliable, affecting functions such as deployment triggers and chat alerts when events arrived late or did not arrive at all.
  • By midday on 17 Aug 2026, GitHub had not identified the underlying cause publicly, declaring it was applying mitigations based on what its investigation had found, and was watching the platform for signs of recovery. Some products—including Git Operations, Packages, Pages, and Codespaces — had remained available, but several central services were still operating in a degraded condition.

The event was at least the second significant Microsoft-platform outage in recent weeks, after an Xbox services disruption late in the previous month. GitHub had also experienced a distinct incident on 6 August 2026 involving Actions, when 71% of workflow runs reportedly failed because of infrastructure problems.