Quay.io Intermittent HTTP 502 on Pull/Push
May 11, 14:12 UTCMay 9, 19:30 UTC
Duration
0m
Impact
Major
Root cause
Not disclosed
Red Hat, 90 days
32 incidents
Affected
Not listed by the vendor.
Final update
On May 9, a surge of concurrent automated image pushes caused database lock contention that degraded Quay.io for approximately 18 minutes. Both image pushes and pulls returned errors during this window. The team identified and disabled an internal feature that amplified the contention, restoring service. We are redesigning this feature to operate asynchronously and improving our database locking mechanisms to better handle high-concurrency traffic.
Timeline
- Postmortem · May 15, 17:11 UTC
On May 9, a surge of concurrent automated image pushes caused database lock contention that degraded Quay.io for approximately 18 minutes. Both image pushes and pulls returned errors during this window. The team identified and disabled an internal feature that amplified the contention, restoring service. We are redesigning this feature to operate asynchronously and improving our database locking mechanisms to better handle high-concurrency traffic.
- Resolved · May 9, 19:30 UTC
For a period of 12 minutes (15:52–16:04 UTC), ~62% of manifest pulls and ~39% of manifest pushes failed with 502s. Quota insights has temporarily been disabled as part of remediation.
More from Red Hat
Full historyAsk Red Hat - Outagelasted 0mQuay.io Push & Pulls Returning Intermittent 500 Issueslasted 2h 32mconsole.redhat.com Notifications - Emails Degradedlasted 4h 1m[Scheduled] Automation Hub Upgradestill open[Scheduled] Automation Hub Upgradestill open[Scheduled] OpenShift Cluster Manager Database Maintenance 2026-09-19lasted 2d 1h