Centralized Monitoring with Munin

As a rule, to solve problems of the server you need to know what happened to it before. Here, monitoring systems come to the rescue. Today there is a wide range of useful monitoring systems (and Nagios is regarded to be the most popular one). Among them Munin should be marked out.

Munin is convenient and easy-to-configure tool of server performance monitoring. It has a large number of useful plug-ins available right away after installing. Its extensible functionality allows to monitor virtually every server characteristic: load average, network, CPU loading, memory usage, postfix, exim4, mysql and many more.

This system consists of two independent parts: the server installed on the machine aimed to collect data (munin itself), and munin-node daemon installed on machines that should be monitored. Daemon itself is a small Perl script that listens to port 4949. Every 5 minutes munin server connects to all nodes to collect information from all plugins. After that it stores received data in own rrdtool database; thus, Munin does not even need MySQL or other database management system.

Incredible ease of implementation allows you to write any plug-in you need at breakneck speed. Probably that is why a relatively young system has already acquired a large number of ready-made plugins.

Extremely useful feature is the existence of munin-node-win that allows to monitor Windows-based servers.

It should be also noted that Munin’s output is easy-to-read html with no similarity to scripts.