Device Failure Hardening

Ideally when a worker hangs, dies and restarts, or dies and fails to restart, we should be able to gracefully handle each case with minimal interruption of service. Need to investigate exactly where we stand on this and add test cases.

Start by reviewing the worker lifecycle and existing test suite to determine how hangs, restarts after death, and failed restarts are currently handled. Add test cases covering each failure mode and verify that service interruption is minimized in the observed behavior.

Written by the indexing model from the issue text.