The problem is that if the connection is disrupted (due to a Redis restart or temporary connectivity issues), the Taskiq worker crashes, interrupting all ongoing handlers. As a result any long-running tasks will be interrupted and will not complete their work.