Maintenance ended

Maintenance is complete. The bad node has been decommissioned. We have successfully failed over to our high-availability standby database and restored normal service operations. We will continue monitoring the system to ensure stability. No data was lost during the incident.

Thank you for your patience and understanding.

At 2026-11-03 01:15:07 UTC, we began experiencing hardware-related failures affecting our primary Postgres database, resulting in an unexpected restart and temporary service disruption. At 2026-11-03 03:14:38 UTC, the affected node experienced a second failure.

To prevent further disruptions, we are initiating an emergency maintenance window to decommission the affected node and fail over to our healthy standby database. During this process, the service may be temporarily unavailable.

No data has been lost as a result of these incidents. Our database remains protected by continuous point-in-time backups, and high availability will be restored following the failover.

We apologize for the disruption and appreciate your patience as we work to restore stability.