Module nagios
In: /modules/nagios/manifests/client/config.pp


# Overview

This module provides a full nagios monitoring solution, for setting up both servers and clients. There can be any number of each, though there is typically one server and many clients. The main limitation is that all server and client nodes need to have all of their nagios related puppet configuration on the same puppetmaster.

The server part is still very Red Hat Enterprise Linux specific. The client part is much more generic and has been tested on both RHEL and Gentoo.

# Requirements

  • Stored configurations enabled on the puppetmaster (mandatory)
  • apache_httpd and php modules used for the nagios server node (optional)

The apache_httpd and php modules are required for the server part, though optionally since it‘s also possible to use an existing non-puppet managed web server or different module modules. For a new empty node, configuring it as a nagios server will be much quicker if those modules can be used.

Stored configurations are essential on the puppetmaster for the module to work at all, since it relies on having all nodes create their own exported nagios host and service resources, which the nagios server node then realizes to build its configuration.

The plugin packages used for RHEL with selinux enabled should have the libexecdir directory properly set to /usr/libexec in order to avoid selinux denials. Such packages can be found on

When SELinux is enforcing, the selinux::audit2allow definition is required to allow some basic nagios/nrpe accesses, though it can also be disabled.

# Sample Usage

Nagios server instance (node specific, typically inside a node section) :

    class { 'nagios::server':
        apache_httpd_ssl             => false,
        apache_httpd_conf_source     => 'puppet:///modules/my/nagios-httpd.conf',
        apache_httpd_htpasswd_source => 'puppet:///modules/my/htpasswd',
        cgi_authorized_for_system_information        => '*',
        cgi_authorized_for_configuration_information => '*',
        cgi_authorized_for_system_commands           => '*',
        cgi_authorized_for_all_services              => '*',
        cgi_authorized_for_all_hosts                 => '*',
        cgi_authorized_for_all_service_commands      => '*',
        cgi_authorized_for_all_host_commands         => '*',
        cgi_default_statusmap_layout                 => '3',

Nagios client instances (typically from site.pp) :

    class { 'nagios::client':
        nrpe_allowed_hosts => ',',

Nagios client specific overrides. See client.pp and check/*.pp for all of the variables which can be manipulated this way. The following :

    nagios::client::config { 'host_address': value => $ipaddress_eth2 }

Will result in having $::nagios_host_name get $ipaddress_eth2 as its value for the entire configuration of the client where it is applied.

Nagios client check override configuration examples :

    nagios::client::config { 'check_ram_ensure': value => 'absent' }
    nagios::client::config { 'check_cpu_args': value => '-w 50 -c 20' }

Configuring a default check (must be done from a scope where ‘nagios::client’ can inherit it) :

    Nagios::Check::Swap { ensure => absent }
    if $::domain == '' {
        Nagios::Check::Cpu { notification_period => 'workhours' }

To enable nagiosgraph on the server :

    class { 'nagios::nagiosgraph':
        # This is the default
        perflog => '/var/log/nagios/service_perfdata.log',
        # To enable the mouseover graphs
        nagios_service_action_url => '/nagiosgraph/cgi-bin/show.cgi?host=$HOSTNAME$&service=$SERVICEDESC$\' onMouseOver=\'showGraphPopup(this)\' onMouseOut=\'hideGraphPopup()\' rel=\'/nagiosgraph/cgi-bin/showgraph.cgi?host=$HOSTNAME$&service=$SERVICEDESC$',
        ssi => true,
    # This is what needs to be changed/added for nagios::server
    class { 'nagios::server':
        process_performance_data => '1',
        service_perfdata_file    => '/var/log/nagios/service_perfdata.log',
        service_perfdata_file_template => '$LASTSERVICECHECK$||$HOSTNAME$||$SERVICEDESC$||$SERVICEOUTPUT$||$SERVICEPERFDATA$',
        service_perfdata_file_processing_interval => '30',
        service_perfdata_file_processing_command => 'process-service-perfdata-nagiosgraph',

To enable nagiosgraph for the client‘s services in the server web interface :

    class { 'nagios::client':
        service_use => 'generic-service,nagiosgraph-service',

# Hints

Debug any startup or configuration problems on the server with :

    nagios -v /etc/nagios/nagios.cfg

A lot can be configured semi-dynamically for nagios::client :

    class { 'nagios::client':
        host_notification_period => $::domain ? {
            /\.dev$/ => 'workhours',
            default  => '24x7',
        # You will need to use the type "nagios_hostgroup" on the server for
        # all of the possible domain values to create the hostgroups.
        host_hostgroups => $::domain,

# Notes

Client overrides should be done in this order :

  • For all services on some nodes : Using nagios::client service_* parameters
  • For all services on a node : Using the nagios::client host_* parameters
  • For specific services on all nodes : Using Nagios::Check::Foo { paramname => ‘value’ } (the scope must be the same or lower than where nagios::client is called from)
  • For specific services on a node : Using nagios::client::config { ‘check_*’: value => ‘foo’ } overrides

Note : nagios::client::config can also be used to override just about anything, though you must take into account that since it‘s done using custom facts, you‘ll need two puppet runs on the client node then one on the server node for the change to be applied : One for the fact to be created, then another for the client configuration to take the new fact into account, then the server run to update the nagios configuration. This might take a little while depending on how often puppet is run on the nodes.


host   service  

Custom Facts

nagios_dns   nagios_moxi   nagios_contactgroups   nagios_httpd   nagios_membase   nagios_couchbase   nagios_hostgroups  

Classes and Modules

Class nagios::check
Class nagios::client
Class nagios::defaultchecks
Class nagios::nagiosgraph
Class nagios::params
Class nagios::server


host( $check_period => 'nagios::client::host_check_period', $use => 'nagios::client::host_use', $contact_groups => 'nagios::client::host_contact_groups', $check_command => 'nagios::client::host_check_command', $host_alias => 'nagios::client::host_alias', $notification_period => 'nagios::client::host_notification_period', $notes => 'nagios::client::host_notes', $address => 'nagios::client::host_address', $server => 'nagios::client::server', $hostgroups => 'nagios::client::host_hostgroups', $notes_url => 'nagios::client::host_notes_url' )

Define: nagios::host

Wrap around the original nagios_host :

  • To be able to export with the default tag
  • To be able to use defaults set by the module

Note: the original nagios_host "alias" parameter is named "host_alias" here since otherwise it‘s interpreted as the puppet "alias" metaparameter.

service( $check_period => 'nagios::client::service_check_period', $use => 'nagios::client::service_use', $max_check_attempts => 'nagios::client::service_max_check_attempts', $contact_groups => 'nagios::client::service_contact_groups', $servicegroups => 'undef', $service_description => 'name', $check_command, $notification_period => 'nagios::client::service_notification_period', $server => 'nagios::client::server', $ensure => 'undef', $host_name => 'nagios::client::host_name' )

Define: nagios::service

Wrap around the original nagios_service

  • To be able to export with the correct tag automatically
  • To be able to use defaults overridden or from facts