The Underscore Archives

					As other replies have said, there are a bunch of apps out there that  
do this.  However, I don't really like any of them.

My problem is that they're all far too complex for my tastes, and the  
code tends to be pretty messy.  When it comes to fault monitoring, I  
want a simple, clean, reliable piece of code doing the job.  I also  
want redundant off-site monitoring, so I can see if the monitoring  
mechanism itself has failed.

I mean, is a monitoring system really useful if you can't trust that  
it's actually reliable?

I never had warm fuzzies over Nagios or Big Brother, and the system I  
was obliged to use at HP was far too byzantine to give us any level  
of confidence.  I ended up putting together a Perl script to do it.

I had a proof-of-concept for a nice clean system back in 1998, where  
a group of independent entities (eg. companies on Underscore) would  
monitor themselves and each other, as an exercise in mutual self- 
interest.  Unfortunately, I didn't have time to develop it further.

Incidentally, using email and SMS for fault notification isn't  
actually something you should rely on, as both can fail silently...  
doubly-so when using an email-to-SMS gateway.  Just this week, I had  
some SMS texts arrive over 8 hours after sending.  That was Orange to  
Orange, with both phones getting a good quality signal.

Tom

-- 
Tom Gidden
http://gidden.net/tom/