As @larsks already commented: from the perspective of a sysadmin you run your process from a suitable supervisor such as systemd or one of the others listed on https://en.wikipedia.org/wiki/Process_supervision
That should take care of the process getting started at system boot, restarting it at failures, stopping the service when requested and some logging.
Second: an external monitor that checks if the system where your software runs and the software itself still works as intended, is healthy, has sufficient resources etc. and alerts you when it isn’t.
As a sysadmin I typically can’t do anymore than that. When your custom application is crap then that is what needs fixing (and redirecting errors to /dev/null is not helpful for detecting errors )