Between 09:42 and 10:19 UTC, approximately 8% of API requests served by the EU edge experienced elevated latency (p95 up to 2.1 s). Root cause was a congested transit provider link; traffic was re-routed and latency returned to baseline. No data loss occurred.

A backlog in the outbound delivery queue delayed webhook deliveries by up to 11 minutes between 14:05 and 14:59 UTC. Deliveries caught up automatically; no events were dropped. A capacity increase has been scheduled to prevent recurrence.

Planned primary database failover for minor version patching. Read-only mode was active for 22 minutes starting 02:00 UTC. All writes queued and were applied after the window. Performed within the published maintenance schedule.