I have a kafka connect cluster with > 500 connectors (custom). At times a chunk of the tasks go into "unassigned" state (> 50) and just stay there. At what conditions does this happen and is there any way to figure out why it stays in unassigned ?
This is on kafka connect version 2.4.1. I saw this When does a task in a connector move to unassigned state? but don't see any reason that would apply here