Scheduled single person timeout

We had some serious trouble with our self-hosted jitsi instance today - people where getting kicked and rejoined. What we see in the logs:

Feb 01, 2022 5:26:36 PM org.jitsi.utils.logging2.LoggerImpl log
INFO: Scheduled single person timeout.
Feb 01, 2022 5:26:36 PM org.jitsi.utils.logging2.LoggerImpl log
INFO: Electing new owner: ChatMember[ticketing@muc.meet.jitsi/5002ce05, jid: null]@1918162832
Feb 01, 2022 5:26:56 PM org.jitsi.utils.logging2.LoggerImpl log
INFO: Timing out single participant: 5002ce05

We run the docker image fee9d08e911a - so v2.0.6826

Some people had a stable experience - others where dropped multiple times - so calls needed to switch to closed external platforms which I really hate to see.
Any idea how to debug this further and avoid that from happening? Anyone else saw these issues on their instances?

This means there was just one participant in the call. Check nginx logs maybe nginx was dropping the connections …