Small Percentage of Egress Failures

Incident Report for LiveKit

Resolved

We received reports of Egress failures on 10/21. We identified a rare condition that caused a few of our Egress instances to not realize they had jobs running when scaling down. A very small percentage of Egresses, around 0.25%, were affected and would have received the error 'egress timed out'. Unfortunately, these recordings are not recoverable due to this issue. We have successfully rolled out a fix and our monitoring has indicated the issue is resolved.

Posted Oct 22, 2024 - 22:00 PDT