Passive Monitoring

I have a host that for specific reason I cannot install the NPRE daemon on it. In order to monitor its internals I have to make it report its status to the Nagios Master.

First, install the following batchcheck.sh script in the target host.

batchcheck.sh

#!/bin/sh

# File: /etc/nagios/batchcheck.sh
# Batch run passive checks

# customize your config here:
NAGIOS_SERVER_IP=10.0.0.1
MY_HOSTNAME=example-host
SEND_NSCA=/usr/bin/send_nsca
SEND_NSCA_CFG=/etc/nagios/send_nsca.cfg
PLUGINS_DIR=/usr/lib/nagios/plugins

for check_cmd in $PLUGINS_DIR/*; do

        svc_description=`basename $check_cmd`
        plugin_output=`$check_cmd`
        return_code=$?
        printf "%s\t%s\t%s\t%s\n" "$MY_HOSTNAME" "$svc_description" "$return_code" "$plugin_output" | $SEND_NSCA -H $NAGIOS_SERVER_IP -c $SEND_NSCA_CFG

done       

Then, put all check scripts in a folder, for example /etc/nagios/checks.

Example for the CPULoad:

# File: /etc/nagios/checks/CPULoad
/usr/lib/nagios/plugins/check_load -w 5.0,4.0,3.0 -c 10.0,6.0,4.0

Add a cron job to run the batchcheck.sh script periodically. Example:

# File: /etc/cron.d/nagios_checks
1-59/5 * * * * nagios /etc/nagios/batchcheck.sh 

Make sure you enabled NSCA (Nagios Service Check Acceptor) at the Nagios master.

Define the template for passive checking:

define service{
        name                            example-passive-generic-service
        ...
        active_checks_enabled           0
        check_freshness                 1
        freshness_threshold             1800
        check_command                   no_report_warn
        ...
        }      

And sample service definition for CPULoad check script:

define service{
        use                             example-passive-generic-service
        host_name                       example-host
        service_description             CPULoad
        ...
        }              

no_report_warn calls a script to set the service status to Warning State. It is used because there is no way to do active check when the freshness_threshold is reached and the system is forced to do a active check. This will happen when for some reason we are not receiving status reports.

#!/bin/bash
# file: /usr/lib/nagios/plugins/no_report_warn.sh

echo "WARNING: Did not receive service status report for a long time!";
exit 1;