Temporary Egress Availability Issues

Incident Report for LiveKit

Resolved

From 15:58-16:18 UTC, around 2% of StartEgress requests failed with 503 service unavailable.

The root cause was an issue with our autoscaling metrics, causing our canary clusters to stop scaling. The change has been reverted and service is now back to normal.

Posted Aug 22, 2025 - 13:00 PDT