Currently, when a delegate undergoes a graceful shutdown, it is not considered a disconnection event and, as a result, does not trigger a notification. However, disconnection events caused by other factors, such as network connectivity issues or pod termination (without graceful shutdown), still trigger notifications as expected.

Regardless of the reason—whether it’s a graceful shutdown or not—we would prefer to receive notifications for all delegate disconnections. Being notified of such events in real-time, whether initiated by code or human intervention, would greatly enhance our ability to monitor infrastructure effectively. We believe it would be valuable for Harness to implement this feature to improve visibility and response time.