The key with nagiosgraph is understanding the map file. This works by using perl regexps to identify what sort of data Nagios has just spat out. You also have to let it know whether you’re getting perfdata or just normal output for the plugin. Looking at one entry in my map as follows
Service type: ntload
check command: check_nt -H Address -v CPULOAD -l5,70,90,30,70,90
output: CPU Load 9% (5 min average) 11% (30 min average)
perfdata: ‘5 min avg Load’=9%;70;80;0;100 ‘30 min avg Load’=11%;70;90;0;100
#/output:.?(d+)% .?(d+)% /
/perfdata:.*5 min avg Load=([0-9]+)%;[0-9]+;[0-9]+;[0-9]+;[0-9]+ 30 min avg Load
and push @s, ntload,
avg05min, GAUGE, $1 ],
avg30min, GAUGE, $2 ] ];
So we’re only looking at perfdata outputted for this plugin. The 1st ([0-9]+)% is put in a guage type rrd called hostname_ntload_avg05min and the second into an rrd called hostname_ntload_avg30min.
As Jakkedup correctly guessed, there’s a lot of help both on the site and on the sourceforge pages. Again, being able to generate new entries for the map file is really the key.
To be honest, if I was in the poisition that you are now, I would go for nagiostat instead. Whilst nagiosgraph is pretty simple to get going, and it improved my understanding of perl regexps you are pretty limited in what you can output. When I get round to migrating my install to the box I’ve been promised for a while now I’ll going to convert to nagiosstat.
Oh and don’t even thing of trying APAN - a whole other world of horror awaits you there