...
Restart Routing: Instead of depending on two hops to route the restart requests, we could distribute the full task assignments to all workers on every rebalance. Then each worker would know exactly where to route the restart request. Unfortunately, this makes the overhead of the rebalance protocol excessive as the number of workers increases (the order of the message size is O(n^2) where n is the number of workers).
Response Codes: 204 on successful restart, 404 if the connector doesn't exist, and 409 if the task is already being restarted