On Wed, 31 Aug 2011, Nathan Ward wrote:
I recently looked in to this, and everything free out
there seems to
suck unless you have people full time building+maintaining this stuff.
Hi Nathan. I think that monitoring applications are inherently labour
intensive. When I've seen them try to do automatic configuration the
results have usually been poor. It can auto detect information on the
systems but can't detect what the organisation considers important or what
thresholds are right for that organisation. You still need to tunr
reporting and alerting to match requirements.
The best way to manage it for medium-large deployments IMHO is to use a
configuration management system like Puppet and have it configure the
monitor for you.
I am very impressed with collectd for performance monitoring and you can
tie it in to Nagios for alert monitoring.
Nagios appears to be the only free tool with almost
dependency behaviors, and flap detection. However it doesn't easily
There is a fork of Nagios called Icinga that may be worth looking at.
Email: robert(a)timetraveller.org Linux counter ID #16440
IRC: Solver (OFTC & Freenode)
Director, Software in the Public Interest (http://spi-inc.org/
Open Source: The revolution that silently changed the world