ImprovedSource - Monitoring your Network

Monitoring your Network
by Cory Rauch 2007-05-06 Category: Linux-System

If you have a bunch of Linux boxes or remotely manage Linux boxes you may want to consider a monitoring solution. Monitoring software can be a lifesaver, especially when you're not physically near the computers in questions. It also provides for smoother service operation by automatically restarting dead daemons even before you may know of a problem. But at the same time some of these software solutions have become quite complicated and expensive to set up. Well in this article we are going to cover a simple to set-up, cheap (free), and flexible monitoring package called 'monit'.

The 'monit' Software Package
The 'monit' package is a daemon monitoring daemon (DMD), written by Jan-Henrik Haukeland, that monitors service daemons continuously to provide both status and daemon restart capability. It also features email alerts, TCP/IP tcp/udp port checking, Protocol verification (HTTP, SMTP, FTP, POP, IMAP, NNTP), Daemon binaries verification via MD5 checksums, and easy configuration.

Installation
Installation is simple: Download the tar-ball and uncompress in a directory, 'cd' over to that directory, and type:

# ./configure; make
# make install

If all goes well it should have installed the 'monit' software to your machine. Repeat for each machine until done.

Next, We need to configure the monitoring done by 'monit'.

Configuration
Monit uses the configuration file '/root/.monitrc' to monitor the local system. The file is pretty straight forward, especially with an example to look at. Below is a sample configuration file for monitoring Apache, SSH, and NAMED.

Sample /root/.monitrc file:

#
# Sysbotz Sample monit configuration.
#

set daemon 300
set logfile /var/log/monit

check apache with pidfile /var/log/httpd/httpd.pid
start = "/root/apache-start"
stop = "/root/apache-stop"
checksum /usr/local/bin/httpd
timeout(3, 3) and alert me@company.com
host foo.com port 80 protocol http
host bar.org port 80 protocol http

check sshd with pidfile /var/run/sshd.pid
start = "/root/sshd-start"
stop = "/root/sshd-stop"
timeout(3, 3) and alert me@company.com
checksum /usr/local/sbin/sshd

check named with pidfile /var/run/named.pid
start = "/root/named-start"
stop = "/root/named-stop"
checksum /usr/local/sbin/named
timeout(3, 3) and alert me@company.com
port 53 use type udp

The first two lines set the time to poll (polling cycle) each service and where to store the log file.

The next lines specify the services to check. All service definitions follow the syntax: check with pidfile [full path to PID file], then optionally following with the start and stop commands to use.

Additional options include:
checksum
This enables the daemon binary verification using MD5 checksums to make sure the daemon binary has not changed during it's running time.

timeout (, ) and alert
This enables the timeout and alert functions which works by alerting the email address of service problem when number of restarts in number of polling cycles occurs. For example, 3 restarts in 3 polling cycles or 3 restarts in 900 seconds (300 second polling cycle) would trigger a email alert notification.

host port protocol
This tells monit to check the specified host/port/protocol combo. An example of use could be to monitor a list of virtual web-sites on a web server.

Conclusion
The 'monit' monitoring package provides a simple and neat way of monitoring your services on your network. Monit provides an extra level of protection from those daemon deaths and helps you during the troubleshooting phase too. For best coverage you may want to install this on all your Linux servers providing important services. Pushchange members can automatically install this package across their Linux system by click the 'Pushchange Install' link below.

Links

Website: Monit Project

Other ImprovedSource Articles:
Boot Fedora Linux Faster: How I Modified Fedora To Boot In Under 25 Seconds
Linux Fastest Boot Time Challenge
How to make a System Restore CD-ROM

[ Back ]