Disruption with viewing some repository pages from large organizations

Severity: Major
Category: Scalability
Service: GitHub

This summary is created by Generative AI and may differ from the actual content.

Overview

GitHub experienced a disruption between October 23, 2025 19:27:29 UTC and October 27, 2025 17:42:42 UTC, where approximately 5,000 logged-in users across less than 1,000 repositories (including forks) encountered timeouts and "unicorn" errors when viewing repository landing pages in the desktop web UI. The impact was limited to repositories within organizations having over 200,000 members. Git operations remained functional throughout the incident. The root cause was feature-flagged changes impacting organization membership, which led to unintended timeouts during membership count evaluations. The issue was resolved by turning off the feature flag and deploying a fix that included optimizations for large organizations, also resolving related timeouts on the Access settings page.

Impact

Approximately 5,000 logged-in users across less than 1,000 repositories, including forked repositories, experienced timeouts when viewing repository landing pages. The impact was specifically limited to users accessing repositories in organizations with more than 200,000 members. Users observed "unicorn" errors in the desktop web UI. Git operations were not affected. Related timeouts on the Access settings page of impacted repositories and forks were also present.

Trigger

The incident was triggered by feature-flagged changes that impacted organization membership. These changes inadvertently caused timeouts during the evaluation of organization membership counts.

Detection

The issue was detected, leading to an investigation. While the specific method of initial detection (e.g., internal monitoring, user reports) is not explicitly stated, the incident progressed to an 'Investigating' status, indicating awareness of the problem.

Resolution

The resolution involved turning off the problematic feature flag. Subsequently, a fix addressing the timeouts was deployed, which also included additional optimizations designed to better support organizations of this size. This deployment also resolved related timeouts on the Access settings page of the affected repositories and forks.

Root Cause

The root cause was feature-flagged changes that were introduced, impacting organization membership. These changes led to unintended timeouts during the evaluation of organization membership counts, which in turn prevented repository landing pages from loading correctly.