snag

Welcome to snag, a package for (remote) monitoring under Nagios.

This package competes primarily with the NRPE mechanism used to perform
checks on remote hosts.  Chief advantages:

* It uses SSH as transport, so it's taking advantage of a package with
robust security, and which you probably have to have running anyway to
admin your servers.  You don't have to open up a separate service for
monitoring.

* (It can use other transports; ultimately check_snag just has to be able
to run something that causes snag to be run on the appropriate host with
standard output coming back.  Netcat, xinetd, or the like could be pressed
into service.  However, SSH is what was intended.)

* You get a free SSH service check out of this.

* You get a command you can run directly on the host as a quick
health-check (especially after making system changes!).

* It runs several checks per session, rather than requiring multiple
sessions to be opened per monitoring cycle.  Most of the standard checks
require a handful of syscalls and can be done more efficiently as a
batch.  

* Since it's enumerating mounted filesystems, it will do the set of checks
appropriate to the particular host.  If /var is a filesystem, you get a
/var check, otherwise not.

* You don't have to configure each check separately; a small Perl script
is included to check for check results coming back and propose Nagios
config file additions.

* You get a single check turning red if the host crashes or becomes
unreachable; it doesn't falsely claim that all the filesystems suddenly
filled up, etc, as the result of failed NRPE checks.

* You don't get several NRPE checks hitting simultaneously and causing a
load spike negatively affecting the host's primary function.  Snag is
single-threaded so its maximum load is limited.

How it works
------------

The snag executable generates output that looks something like this:

Disk space on /;1;DISK WARNING - / 2578 MB (19%) free space|/=2578MB
Disk inodes on /;0;INODES OK - / 1052605 inodes (59%) free|/=1052605 inodes
Disk space on /home;1;DISK WARNING - /home 9953 MB (19%) free space|/home=9953MB
Disk inodes on /home;0;INODES OK - /home 6458748 inodes (98%) free|/home=6458748 inodes
Load average;0;LOAD OK - load average: 0.57 0.55 0.40|load1=0.57 load5=0.55 load15=0.40
Uptime;0;UPTIME OK - up 21 days|uptime=1825881s
Swap free;0;SWAP OK - 99% of swap free|swap=99
Processes;0;PROCS OK - 148 total processes|procs=148

There is a config file that can be used to tune thresholds for these
standard checks, and to add Nagios Plugin checks.

The check_snag executable runs on the Nagios server, configured as a
normal Nagios active check.  It gets run by Nagios at the configured
schedule, and reports OK, WARNING, CRITICAL depending on what happened
when it tried to run snag on the target host.  Check results it got back
are fed to the Nagios external command pipe as passive check results.

Installation
------------

See the file INSTALL in the distribution.

Author/Copyright
----------------

Copyright (C) 2006 Anthony de Boer, released under GPL v2.