check_netsnmp plugin for Nagios

Motivation. There are several common parameters we want to monitor for (almost) every computer in our network:

• Free disk space
• Free swap space
• Average loads
• That important processes are running on the machine
• Total number of processes and a number of processes in zombie state
• A number of free mbuf clusters available (most our servers run FreeBSD, thus the notion of mbuf clusters)

check_netsnmp plugin for Nagios is written in perl, uses no external commands and requires utils module, available from Naguios-plugins distribution, and Net::SNMP module, available from CPAN.

Usage. Put check_netsnmp plugin to the directory in which you installed nagios-plugins (typically $USER1$ variable points to it).

1. Monitoring free disk space. Used in conjunction with disk keyword in snmpd.conf. Imagine that you have two mount points on the monitored machine: / and /var and you want to be sure / has more than 20% free space and /var has more that 30% free space. Put the following lines to the snmpd.conf:

   disk / 20%
   disk /var 30%
   

   Next, define the following command in your Nagios configuration:

   define command{
           command_name    check_disk
           command_line    $USER1$/check_netsnmp -H $HOSTADDRESS$ -o disk
   }
   

   and add the relevant service:

   define service{
           use                     generic-service
           host_name               host.name.ru
           service_description     disks
           check_command           check_disk
   }
   

   The same procedure applies to checking swap, processes and load averages (use with swap, proc and load keywords in snmpd.conf).
2. Monitoring free mbuf clusters. This is a bit harder since there are no corresponding mibs in snmpd. Here is where snmpd's exec keyword may be used. Write your own script which checks for mbuf clusters and returns 0 if there are enough of them free, and 1 if there are not. An example may be found here. Then put the folowing line in your snmpd.conf:

   exec .1.3.6.1.4.1.2021.8.3 MBufMon /usr/local/share/snmp/exec/mbufmon.sh
   

   Define the following command in your Nagios configuration:

   define command{
           command_name    check_mbufmon
           command_line    $USER1$/check_netsnmp -H $HOSTADDRESS$ -o 3
   }
   

   You may use the full OID instead of just 3 here, if you export the result of mbufmon.sh under nonstandard OID. .1.3.6.1.4.1.2021.8. is prepended to 3 in this example automatically. Finally, add the service to check for mbufs:

   define service{
           use                     generic-service
           host_name               host.name.ru
           service_description     MBufMon
           check_command           check_mbufmon
   }