I don't really care about windows workstations being down. Perhaps a warning after 1 week downtime would be nice.
On the other hand, while it's up I would like alarms from other monitors to get handled. If cpu load goes critical i want a critical alarm. if cpu load fails because host is down i don't want any alarm at all.
Has anyone done this?
As a first step I have defined a command called exists_dns and use that as the check_command for windows workstations. This way a host is considered up as long as it exists in dns. I still don't like having the ping service flagged as critical when people go home, but at least i don't have a bunch of hosts down to clutter up the interface.