Getting those messages is very different (and a lot more reasonable) from forceful restarting them, which was the initial suggestion.

The restarting risk is normally so small, that there are several other things that are more important than constantly restarting to test that restart works. Continuous delivery, security patching and hardware failure will likely cause enough restarts anyway.