NRPE Handler

An event handler designed to handle remote host and service issues via NRPE.

Download nrpe_handler

NAME

nrpe_handler.pl

SYNOPSIS

 nrpe_handler.pl [OPTIONS] --cmd=DOWN:hostname:command:arg1,arg2,arg3         $HOSTSTATE$ $HOSTSTATETYPE$ $HOSTATTEMPT$ $HOSTDOWNTIME$
 
 nrpe_handler.pl [OPTIONS] --cmd=CRITICAL:hostname:command:arg1,arg2,arg3         $SERVICESTATE$ $SERVICESTATETYPE$ $SERVICEATTEMPT$ $SERVICEDOWNTIME$

DESCRIPTION

Nrpe_handler is a Nagios event handler designed to execute programs via NRPE. It is keyed around host and service state to ensure maximum flexiblity handling host and service problems.

There are no requirements for installation on the Nagios host, other than Sys::Syslog if that is desired. Edge machines will need a working NRPE daemon running typically on port 5666. This will likely require tweaks of the nrpe.cfg file and potentially changes to /etc/sudoers if elevated access is required. See the necessary docs. (hint in sudoers: turn off requiretty)

OPTIONS

-c, --cmd=STATE:HOST:COMMAND:ARG1,ARG2,ARG3

Command definition. STATE: in what hoststate or servicestate to run the command HOST: on which host to run the command COMMAND: command to run via NRPE on the remote host ARGLIST: optional comma-separated list of arguments to pass to the command

More than one command can be run by issuing multiple --cmd options.

-d, --debug

Print debug info to logs, if logging.

-l, --log=LOGFILE

Log messages to given file.

-s, --syslog

Log messages to syslog.

$STATE$ $STATETYPE$ $ATTEMPT$ $DOWNTIME$

Nagios macro variables. Host handler: hoststate hoststatetype hostattempt hostdowntime. Service handler: servicestate servicestatetype serviceattempt servicedowntime.

CONFIGURATION

Sample configurations for host and service checks.

Host Handler

Assume we are using machine nagios1 to monitor a host named virt1, a VM running on blade1. We ultimately want to restart virt1 when it goes down by connecting to blade1 and issuing the necessary commands.

On nagios1, Nagios config:

 define host{
        use             generic-host
        name            virt1
        check_command   check-vm-alive
        event_handler   restart_vm!blade1!virt1
        }
 
 define command{
        command_name    restart_vm
        command_line    $USER1$/nrpe_handler.pl --debug --log=/tmp/restart_vm.debug 
                                        --cmd=DOWN:$ARG1$:restart_vm:$ARG2$ 
                                        $HOSTSTATE$ $HOSTSTATETYPE$ $HOSTATTEMPT$ $HOSTDOWNTIME$
        }

On blade1, NRPE config:

 command[restart_vm]=/usr/lib/nagios/plugins/restart_vm.sh $ARG1$

Service Handler

Assume nagios1 monitors a temperamental LDAP service on ldap1. We want to restart that service when it goes critical by connecting to ldap1 and issuing the necessary commands.

On nagios1, Nagios config:

 define service{
        use                             generic-service
        service_description             LDAP Service
        host_name                       ldap1
        check_command                   check_ldap
        event_handler                   fix_slapd!ldap1
        }
 
 define command{
        command_name    fix_slapd
        command_line    $USER1$/nrpe_handler.pl --syslog 
                                        --cmd=CRITICAL:$ARG1$:fix_slapd 
                                        $SERVICESTATE$ $SERVICESTATETYPE$ $SERVICEATTEMPT$ $SERVICEDOWNTIME$
        }

On ldap1, NRPE config:

 command[fix_slapd]=/usr/lib/nagios/plugins/fix_slapd.sh

CHANGES

nrpe_handler 1.00 (20120626)

- Initial release.