I recently updated a Iarge cluster of 600+ postgres connectors from a 1.x beta release to the latest 1.1 stable. I immediately started to run into issues with the connector tasks failing and the tasks having issues restarting due to the associated replication slot still being in use. Things were so bad I had to downgrade back to the beta. I believe I have tracked down the issue to a refactor of the base task state management / coordinator refactor.

My apologies if this ticket is not in an aideal format, I am making it just so that I have something to tie the PR to.