Monitoring server via snmp?

[Brent Bice]

John Heim wrote:
> Thanks to all who helped me with my question about monitoring a server via 
> snmp.  I have 2 more questions:
> 
> First, can I check the load on the server? 

    I use nagios for monitoring the health of servers/services and 
things I want to be paged (or emailed) on, but I use something else for 
monitoring server/service performance. I treat performance monitoring 
and up/down/degraded sort of monitoring separately because I usually 
want to track performance stats for far more things than I want to be 
paged about. (grin)

    For performance monitoring, my favorite tools are net-snmp and 
cacti. I instrument the servers with net-snmp and then monitor their 
performance stats with cacti. This way I can look at stats like cpu, 
load average, memory usage, swap usage, swap I/O, disk I/O stats, 
filesystem usage, etc.

    net-snmp is very extensible and you can hook all sorts of other 
things into it too. For instance, I have a perl script hooked into it 
that monitors postfix stats like messages/bytes sent/received, queue 
sizes, etc. And cacti can then graph these stats too. In fact, several 
years ago I wrote an article for SysAdmin magazine on net-snmp that I 
could probably dig up and re-post if anyone was interested.

    The nice thing about monitoring with cacti/rrdtool is that you can 
not only look at a machines stats right now, but look for trends over 
longer periods.  It's pretty slick.

http://www.cacti.net/
http://www.net-snmp.org/

Brent
bbice at sgi.com