Several weeks ago, we started using Ganglia to monitor all our servers. And I’m glad I did. For years we used a home-made system built on sysstat and RRDTool and it served its purpose, but then I discovered Ganglia. Our old system gave us performance statistics on a 10-minute delay but Ganglia is almost live, which is very helpful to see what’s happening now. It’s also much easier to setup, just apt-get ganglia-monitor, edit the /etc/gmond.conf file (if necessary) and a new server is being monitored. It’s that easy (if you use Debian).
Ganglia was originally funded by the place I interned for while in college (and, not coincidentally, my dad’s employer for 25 years), the National Science Foundation. That’s good government at work! :)