Introduction

Monitorix is a lightweight system monitoring tool that includes monitoring from the CPU load and temperatures to the users using the system. Network device activity, network services demand and even the devices' interrupt activity are also monitored. It uses the Apache web server to show the graphs.

This tool was designed primarily to be used on Linux servers, but of course you can use it on your laptop or even on your Linux box at home.

Since March 2006, Monitorix has a minimal support to run on FreeBSD systems. My special thanks to twenty4help Knowledge Service (http://www.twenty4help.com) and to Roger "Rocky" Vetterberg for their support, help and for to be a good friends during all the portability process.

I use RedHat/Fedora/CentOS Linux distributions, so this project was made it keeping in mind these distributions. It is possible you have some problems with other distributions. If so, contact with me. But since the 0.8.0 version, it should be easier to port Monitorix to other Linux distributions and even to other *NIX systems.

Besides the above mentioned Linux distributions, there are supported too the following ones. All they are covered by the install.sh script:

  • Debian
  • Gentoo
  • Slackware

For people coming from other Linux distributions the same install.sh comes with the Generic option. You can adjust the paths before using it.

Monitorix has a configuration file located in /etc/monitorix.conf to adapt your server to it. Every configuration option is commented in the same file, although for a complete description you can read the file Configuration.help included in the package.

NOTE: The configuration file is a Perl file where you set the value directly to the variables that Monitorix will use during its normal operation. So you must take special precaution to not break some Perl basic syntax. I mean, every variable need to terminate with a semicolon ";".

Please, see the Configuration.help file.

Requirements

This package requires some others packages to be installed that your Linux distribution may not have them:

Installation on a RedHat/Fedora/CentOS Linux

Once downloaded the RPM file, you can install it using: 
# rpm -ivh monitorix-n.n.n.noarch.rpm
(where n.n.n is the latest version)
Once succesfully installed, you may want to make some adaptations in the configuration file located in /etc/monitorix.conf. And finally you'll be able to start the Monitorix shell-script based daemon with: 
# service monitorix start
At this point, Monitorix will start collecting system information, and you'll be able to see it from your favorite browser pointing it to:

http://localhost/monitorix

Installation on a generic UNIX/Linux system

Once downloaded the tar.gz file, you can install it using: 

# cd monitorix-n.n.n
(where n.n.n is the latest version)

# ./install.sh
Once succesfully installed, you may want to make some adaptations in the configuration file located in /usr/local/etc/monitorix.conf. And finally you'll be able to start the Monitorix shell-script based daemon with: 
# /usr/local/etc/rc.d/monitorix.sh start
At this point, Monitorix will start collecting system information, and you'll be able to see it from your favorite browser pointing it to:

http://localhost/monitorix

Notes after installation

If you are redirecting all the output to /var/log/syslog, you may want to add the cron.none directive into the /etc/syslog.conf, in order to avoid annoying messages from crond about Monitorix.

If your system is a Fedora Core 3 or newer, so you have probably SELinux enabled by default, it is possible your Monitorix graphs are not visible and you get lot of messages in dmesg and in /var/log/messages about access denied to the RRD database files. I don't know enough SELinux to adapt Monitorix to it, so in that case my only recommendation is to disable SELinux, adding the line SELINUX=disabled in /etc/selinux/config or with the kernel parameter selinux=0.

If you stops the Monitorix service and after long time you restarts it, you may notice that your network graphs show a very high load (peak). This is because, although the cron has been stopped, the iptables chains continues processing the counting, and then when the cron restarts, Monitorix finds a value much greater than when it had stopped.

Check that the /var/log/httpd directory have enough read permissions for the user running the Apache server. In the newer Fedora Core versions this directory have the permissions restricted to only the root user.