After a service rollout, a replacement migration job started but did not progress. Investigation found an older database operation still holding a lock, even though the service task that started it was gone. Restarting the application had not ended every piece of work it had launched.
The project update in Duva recorded the waiting operation, the stale blocker, the intervention, and the fact that the replacement job moved forward. It also linked the follow-up code fix and explained why another rollout needed to wait. That kept the immediate recovery separate from the follow-up engineering work.
Record what the next worker is waiting for. A stalled-job handoff should identify the dependency, the evidence used to identify it, the authorized action taken, and the next verification. Keep the diagnosis with the incident record so a new session can understand why another restart may not be the next useful step.