From what I’ve read, it’s non-trivial to catch all stale pid cases through process inspection