Setting something as a master is the only way to have it check in every 30 seconds instead of every 5 minutes. Actually, as far as I'm aware, the only thing the master checkbox does is to make it check in more frequently. If you want servers to be quickly accessible and to get notices as fast as possible, it's very useful to have them all set as masters. It was actually a support tech's recommendation that first made us set all our servers as masters. We haven't had any performance or bandwidth issues.
With the 30 second checkin time, we often know internet or servers are down before our clients do (and certainly before they've had a chance to contact us) - a very important thing.
As for the false reports when the internet is down...that's something we deal with too. I've managed to get it to not report when *our* internet is down, but there isn't a reliable way as far as I can tell to avoid false alerts when the internet at outside locations is down. Just last night, we were flooded with **1,000** emails due to some widespread rolling power blackouts.
I guess we could set up a ping monitor against the external interface of all our location's firewalls and then check against that... after the flood of alerts last night I'm tempted to go to the trouble.