a primary VPS that provides web hosting to the public.
a couple small VPSes. I'd like to use either one or both to monitor the primary VPS.
the small VPSes are 192MB/384MB Linux boxen. I'd prefer to run CentOS on them so everything is standardized but that's not 100% required. Nothing else of significance will run on them.
primary and small VPSes are in the same city though different facilities.
I'd like the monitoring solution to make sure something is really down before it emails my phone/email - perhaps monitor at interval X, and if there is a problem, monitor again 60 or 120 seconds later, and only then page. Or something like that - I want to avoid being bugged at 2:30am with transient network problems
It would be nice if a log was kept as well.
I'd also like to do some monitoring on the primary VPS - i.e., if its apache goes down or there is some error in the logs, it should tell me.
A friend recommended Nagios. I'm only familiar with godawful giant monitoring frameworks we use at my day job. Is that the way to go or should I also look at something else?
Shoestring operation so FOSS really preferred but if there's something mind-blowingly industry-standard fantastic that is pay I could consider it.
My advice and reviews of VPS providers based on my personal experience: VPSadvice.com