dead-man-switch

You prevent long-running processes from getting stuck by monitoring for inactivity and triggering a state save plus notification if no progress is detected within a defined timeout threshold. This ensures you can resume operations from where they left off.

A dead-man-switch for monitoring process inactivity works by requiring a process to emit progress signals. If no signal is received within the defined timeout duration, it automatically saves the current state and sends an alert.

You automatically save state and get notified by setting an inactivity threshold for your multi-phase job. The switch detects when progress signals stop, saves the current state for later resumption, and alerts you to the interruption.

Yes, you need a mechanism to emit progress signals to use automated inactivity monitoring. The system relies on tracking these signals against a defined inactivity timeout duration to detect when a process is potentially hung.

You should use an inactivity timeout for complex, multi-step data processing jobs that might take hours. It is specifically applicable to long-running, multi-phase operations where process interruption can lead to significant rework.

A limitation of using a dead-man-switch for error handling is that it only detects inactivity beyond a defined threshold. It requires the process to actively emit progress signals and cannot prevent the initial hang, only save state and notify after the fact.