Supabase incident

pgBouncer issues on some older projects

Minor Resolved View vendor source →

Supabase experienced a minor incident on August 26, 2026 affecting Connection Pooler and Storage, lasting 22h 34m. The incident has been resolved; the full update timeline is below.

Started
Aug 26, 2026, 07:49 PM UTC
Resolved
Aug 27, 2026, 06:24 PM UTC
Duration
22h 34m
Detected by Pingoru
Aug 26, 2026, 07:49 PM UTC

Affected components

Connection PoolerStorage

Update timeline

  1. investigating Aug 26, 2026, 07:49 PM UTC

    Some users on older Projects are experiencing a pgBouncer configuration issue causing connection issues for the Dedicated Pooler and Storage services. This issue should only affect users with Project versions 15.1 and older. We have determined the cause of this issue and we are rolling out a fix.

  2. investigating Aug 26, 2026, 09:36 PM UTC

    Some users on Project versions 15.1 and older are experiencing a pgBouncer configuration issue causing connection issues for the Dedicated Pooler and Storage services. We are continuing our investigation of this issue.

  3. identified Aug 27, 2026, 01:21 AM UTC

    This issue is fixed in eu-central-2 and ap-south-1. We are continuing to rollout the fix to all regions.

  4. identified Aug 27, 2026, 10:12 AM UTC

    We are continuing to rollout the fix to further regions.

  5. identified Aug 27, 2026, 01:04 PM UTC

    An enhanced automated solution is currently being tested in staging. It is expected to be more effective and efficient than the approach used previously. Once testing has completed successfully, we will use this approach to roll out the fix in production. We will provide further updates as this work progresses.

  6. monitoring Aug 27, 2026, 04:34 PM UTC

    A fix has been rolled out across all regions, and results are positive. We will continue monitoring to ensure the service remains stable. Moving this incident to monitoring.

  7. resolved Aug 27, 2026, 06:24 PM UTC

    The fix has remained stable across all regions, and the issue has been resolved.