you dont get it.
if you want to report something is not ok youll need another service which report it in case it does not recieve an ok
so you need an additional service than, more resources more point of errors
reporting 1 or 2 times a minute its ok its the simplest way and compared to the size of logfiles it doenst make any real difference in terms of diskusage or io load
you can easy config a service which reboots in case there is no ok within the last 100 seconds or so - you could even make a simple script todo that
you can never report 100%tly something is NoT working but you can that it is - so its really not useless
and those few lines - man where talking about 120 lines an hour
even oyu dont want to have such a autorestart service it helps nicly to determine when exactly something has crashed - narrow down the source of error
i really dont see a big issue here, and its one seperated logfile for that purpose so what ?