Let me introduce you Zabbix, a very nice all-in-one open source availability and performance monitoring solution that I use everyday on TextMe to ensure that all the applications and servers are going well. Below is a short list of features available in Zabbix:
auto-discovery of servers and network devices (great for autoscaled platforms) low-level discovery (the killer feature of the 2.0 branch, imho) distributed monitoring with centralized web administration support for both polling and trapping mechanisms server software for Linux, Solaris, HP-UX, AIX, FreeBSD, OpenBSD, OS X native high performance agents (client software for Linux, Solaris, HP-UX, AIX, FreeBSD, OpenBSD, OS X, Tru64/OSF1, Windows NT4.0, Windows 2000, Windows 2003, Windows XP, Windows Vista) agent-less monitoring secure user authentication flexible user permissions web-based interface flexible e-mail notification of predefined events high-level (business) view of monitored resources audit log… ...