I'm monitoring a service via check_nrpe to hundreds of boxes. Whenever this service goes bad, typically every box goes bad and get flooded with notifications/text msgs.
Any suggestions on how or what to use to roll all this up in to a single check?
Seems like this would be a good case to utilize the Nagios dependency system. Make all the hundreds of checks dependent on this one check_nrpe, and that should suppress all the extraneous notifications.
In addition to the other fine suggestions given, if you really just literally want to create a new additional check that fires when any of a set of checks fires, try check_service.pl http://exchange.nagios.org/directory/Plugins/Others/Check-many-services/details.
If that's the case first make sure you're monitoring the right thing. If for instance the problem is that a switch fails and all you're doing is monitoring the hosts attached to it you'll get a bunch of host down alerts. Second use dependencies. This allows nagios to notify you on the switch failure and ignore all the hosts down alerts behind it.