This summary is created by Generative AI and may differ from the actual content.
Overview
On August 19, 2025, between 13:35 UTC and 14:33 UTC, GitHub search experienced a degraded state, affecting pull requests, issues, and workflow runs. Users encountered slow, empty, or incomplete search results, and some pull requests failed to load. The incident was caused by intermittent connectivity issues between load balancers and search hosts, which led to retry queues overwhelming the load balancers after initial masking by retry logic. Mitigation was achieved by throttling the search index pipeline.Impact
GitHub search was in a degraded state for 58 minutes, causing slow, empty, or incomplete results for pull requests, issues, and workflow runs. Some pull requests failed to load. The incident affected the Issues and Actions services.Trigger
The incident was triggered by intermittent connectivity issues occurring between GitHub's load balancers and search hosts.Detection
Awareness of the incident was primarily through automated alerting and internal retry mechanisms, which also helped reduce the overall impact.Resolution
The incident was resolved by mitigating the slowness at 14:33 UTC through throttling the search index pipeline.Root Cause
The root cause was intermittent connectivity issues between the load balancers and search hosts. While initial retry logic masked these problems, the continuous retries eventually led to retry queues overwhelming the load balancers, causing the system to fail.
