Between 2025-09-27 19:31:22 PDT and 2025-09-27 20:14:00 PDT, the secondary database node in Prod3 experienced a downtime. The issue was mitigated by restarting the affected nodes.
A maintenance activity aimed at reducing fragmentation and stale state buildup inadvertently caused resource pressure on the system.
Customers were unable to run pipelines during the outage.
Immediate: Restarted the nodes and rolled back recent changes to restore service.
Permanent: Ongoing improvements to prevent recurrence.