Struck in red is what Red Hat (incl. Quay.io) removed; blue is what they added. Muted text is unchanged context.
BeforeAfter
AddedTitle#1
Before: not present
After: Quay.io Push/Pull Degraded
AddedImpact#2
Before: not present
After: major
AddedStatus#3
Before: not present
After: postmortem
AddedStart time#4
Before: not present
After: 2026-04-21 22:50 UTC
AddedResolved time#5
Before: not present
After: 2026-04-22 00:09 UTC
AddedAffected components#6
Before: not present
After: API, Quay.io - API
AddedUpdateinvestigating#7
Before: not present
After: We are experiencing degraded performance on quay.io push/pull API. We are actively investigating.
AddedUpdateidentified#8
Before: not present
After: Pull API is working again. We are actively working on bringing push API back online.
AddedUpdateresolved#9
Before: not present
After: Pull and Push APIs are fully up again. Service is working as expected.
AddedUpdatepostmortem#10
Before: not present
After: On April 21, a routine production deployment included a dependency upgrade that exposed a latent incompatibility, causing Quay.io application pods to crash on startup. Image pulls and pushes were fully unavailable for approximately 12 minutes before the team failed over to a secondary region, restoring pulls. Pushes remained unavailable for approximately 70 additional minutes while the deployment was reverted and traffic was incrementally shifted back to the primary region. We are enabling staging parity for all runtime features, improving automated rollback mechanisms, and isolating observability subsystems from the critical request path to prevent recurrence.
Rewindex commits each version of an incident to a daily checkpoint, chains it to the day before, and has the checkpoint timestamped by FreeTSA and DigiCert. That proves these records existed by then and have not been changed since, by anyone, Rewindex included.