Summary

In rare occasions for a Managed Synchrony setup, it's possible for a Synchrony node to fail sending a heartbeat in time, causing the node to be evicted from the cluster. This node does not rejoin the cluster until an admin restarts Synchrony (through the Collaborative Editing admin page) on that particular node.

Suggestion

Let the affected Synchrony node auto-restart to successfully join the cluster as soon as the node is evicted. When that is done, notify the admin of the incident so it can be looked into.