Pull Requests and Issues unavailable for signed-out users
This summary is created by Generative AI and may differ from the actual content.
Overview
On June 8, 2026, between 06:30 UTC and 08:36 UTC, unauthenticated (signed‑out) users experienced a sustained increase in HTTP 504 gateway‑timeout errors when accessing Pull Requests, Issues, releases, patch diffs, and related GitHub.com pages. Approximately 17 % of unauthenticated requests returned 504 errors, peaking at about 34 % around 06:50 UTC. Signed‑in users were not affected. The degradation was caused by a large surge of abusive traffic targeting those endpoints, which overloaded the request queues beyond timeout limits. Targeted blocking of the abusive traffic at the load‑balancer and application layers restored normal error rates by 08:36 UTC. Future work includes improving automated detection and blocking of such traffic patterns, enhancing the emergency traffic‑blocking deployment path, and evaluating routing changes for endpoints used by both signed‑out users and automated workflows.
Impact
The incident lasted roughly two hours and affected only unauthenticated traffic. Error rates for unauthenticated requests rose to 17 % on average and 34 % at peak, causing many users to see gateway‑timeout pages when accessing Pull Requests, Issues, releases, and patch diffs. Some GitHub Actions workflows that depended on release downloads or related endpoints also experienced degraded performance. Signed‑in users experienced no impact. No direct revenue impact was reported, but the outage reduced availability for a subset of users and automated workflows.
Trigger
A significant increase in abusive traffic directed at specific GitHub.com endpoints triggered the incident. The abusive traffic overwhelmed the service’s ability to process unauthenticated requests, causing request queues to exceed timeout thresholds and resulting in HTTP 504 errors.
Detection
Monitoring systems observed a rapid rise in HTTP 504 errors for unauthenticated requests to the affected endpoints. Alerts were generated, and the on‑call team began investigating the degraded performance, confirming that the issue was limited to signed‑out users.
Resolution
The incident was mitigated by identifying the anomalous traffic pattern and applying targeted blocks at both the load‑balancer and application layers to stop the abusive requests. After the blocks were in place, error rates returned to normal and all affected services were fully restored by 08:36 UTC. Ongoing monitoring ensures stability.
Root Cause
The root cause was a surge of abusive traffic to the affected GitHub.com endpoints, which exceeded the capacity of the request‑handling pipeline for unauthenticated users. The existing traffic‑filtering mechanisms did not detect or block this pattern quickly enough, allowing queues to grow past timeout limits and produce HTTP 504 errors.