Databases Reference
In-Depth Information
Open Source Monitoring Tools
The following are some of the most popular open source all-in-one monitoring systems:
Nagios
Nagios (
http://www.nagios.org
)
is probably the most popular open source problem
detection and alerting system. It periodically checks services you define and com-
pares the results to default or explicit thresholds. If the results are outside the limits,
Nagios can execute a program and/or alert someone to the trouble. Nagios's con-
tact and alert system lets you escalate alerts to different contacts, change alerts or
send them to different places depending on the time of day and other conditions,
and honor scheduled downtime. Nagios also understands dependencies between
services, so it won't bother you about a MySQL instance being down when it no-
tices the server is unreachable because a router in the middle is down, or when it
finds that the host server itself is down.
Nagios can run any executable file as a plugin, provided it accepts the right argu-
ments and gives the right output. As a result, Nagios plugins exist in many lan-
guages, including the shell, Perl, Python, Ruby, and other scripting languages. And
if you can't find a plugin that does exactly what you need, it's simple to create your
own. A plugin just needs to accept standard arguments, exit with an appropriate
status, and optionally print output for Nagios to capture.
Nagios has some serious shortcomings, though. Even after you've learned it well,
it is hard to maintain. It also keeps its entire configuration in files, instead of a
database. The files have a special syntax that is easy to get wrong, and they are
labor-intensive to modify as your systems grow and evolve. Nagios is not very
extensible; you can write monitoring plugins easily, but that's about all you can
do. Finally, its graphing, trending, and visualization capabilities are limited. Nagios
can store some performance and other data in a MySQL server and generate graphs
from it, but not as flexibly as some other systems. All of these problems are made
worse by politics. Nagios has been forked at least twice due to the real or perceived
difficulties of working with the code and people involved. The forks are named
Opsview (
http://www.opsview.com
)
and Icinga (
http://www.icinga.org
)
. Many peo-
ple prefer these systems to Nagios.
There are several books devoted to Nagios; we like Wolfgang Barth's
Nagios System
and Network Monitoring
(No Starch Press).
Zabbix
Zabbix is a full-featured system for monitoring and metrics collection. For exam-
ple, it stores all configuration and other data in a database, not in configuration
files. It also stores more types of data than Nagios and can thus generate better
trending and history reports. Its network graphing and visualization capabilities
are superior to Nagios's, and many people find it easier to configure, more flexible,
and more scalable. See
http://www.zabbix.com
for more information.
