Has anyone seen issues with SQS event replication in aws workers? We have seen: • Workers losing the websocket connection to prefect and logging that at the warning level after which the queue grows until the worker is restarted • The queue growing without any obvious errors / log messages and a failure to dequeue anything on worker restart

I'm aware of at least one other user who's seen this. Can you share the logs you're seeing from your ECS worker?

sure - do you mind a DM for that?

Sure!