ShardingCoordinators generally have the requirement to make forward progress. If there is a bug, a coordinator might get trapped in a retry loop forever, typically stalling the test until the test itself times out. These situations are pretty hard to debug and search in BuildBaron, because the smoking gun is not prominently displayed in the test log, we only have a timeout. This ticket proposes adding a test-only tassert (probably gated behind the testing proctor) that would trigger when a certain timeout or max retry count is reached, crashing the server with a meaningful error message.