Snap with watchdog-timeout enters failed state


I’ve been investigating the issue where one of services in our snap does not recover after being killed by watchdog.

err systemd Watchdog timeout (limit 15min)!
notice systemd Main process exited, code=killed, status=6/ABRT
notice systemd Unit entered failed state.
warning systemd Failed with result 'signal'.

This is the code in the snapcraft.yaml:

    command: usr/bin/
    daemon: simple
    restart-condition: always
    restart-delay: 60s
    watchdog-timeout: 15m

What are the preconditions for it to enter failed state? How can we make sure it always restarts?