We use Nagios for sending/escalating alerts, and Ganglia for graphing quantifiable systems trends. It's indispensable for explaining odd performance phenomena.

We use Nagios for sending/escalating alerts, and Ganglia for graphing quantifiable systems trends. It's indispensable for explaining odd performance phenomena. Ganglia Demo: http://monitor.millennium.berkeley.edu/ http://monitor.millennium.berkeley.edu/ We also wrote some daemons/libs to make it easy to add our own metrics into the Ganglia graphs, so we can see connected users, requests per second, revenue, etc. along side CPU, memory usage, disk space, bandwidth. https://spideroak.com/code https://spideroak.com/code