In August, GitHub experienced a single incident causing degraded service performance.
August 14 Incident (23:02 UTC, lasting 36 minutes)
Between 23:02 and 23:38 UTC on August 14, all GitHub.com services became inaccessible to users.
Root Cause: At 22:59 UTC, an incorrect database configuration deployed to GitHub.com systems. This change prevented databases from responding properly to health verification signals sent by the routing service. Consequently, database hosts were marked as unhealthy, rendering the production read-only database endpoint unreachable. The application lost connection to essential data required for read operations, triggering widespread GitHub.com disruption beginning at 23:02 UTC. The company confirmed no data loss or corruption occurred.
Resolution: The team reverted the problematic configuration change and restored database connectivity. Service traffic resumed at 23:38 UTC with full system recovery. Additional monitoring continued until 00:30 UTC on August 15 before incident closure.
Preventive Measures: GitHub implemented new safeguards in database change management procedures and prioritized enhancement initiatives including faster rollback functionality and strengthened resilience against dependency failures, with these improvements receiving highest priority status.