Icinga Object Configuration

Icinga Network Monitoring

November 2013


For the most complex and dispersed networks, Icinga is the monitoring tool that puts you in control. This book gives you a thorough grounding and improves your capability in using Icinga to extract precisely the reports you need.

(For more resources related to this topic, see here.)

A localhost monitoring setup

Let us take a close look at our current setup, which we created, for monitoring a localhost. Icinga by default comes with object configuration for a localhost. The object configuration files are inside /etc/icinga/objects for default installations.

$ ls /etc/icinga/objects commands.cfg notifications.cfg templates.cfg contacts.cfg printer.cfg timeperiods.cfg localhost.cfg switch.cfg windows.cfg

There are several configuration files with object definitions. Together, these object definitions define the monitoring setup for monitoring some services on a localhost.

Let's first look at localhost.cfg, which has most of the relevant configuration. We have a host definition:

define host{ use linux-server host_name localhost alias localhost address }

The preceding object block defines one object, that is, the host that we want to monitor, with details such as the hostname, alias for the host, and the address of the server—which is optional, but is useful when you don't have DNS record for the hostname. We have a localhost host object defined in Icinga with the preceding object configuration.

The localhost.cfg file also has a hostgroup defined which is as follows:

define hostgroup { hostgroup_name linux-servers alias Linux Servers members localhost // host_name of the host object }

The preceding object defines a hostgroup with only one member, localhost, which we will extend later to include more hosts. The members directive specifies the host members of the hostgroup. The value of this directive refers to the value of the host_name directive in the host definitions. It can be a comma-separated list of several hostnames. There is also a directive called hostgroups in the host object, where you can give a comma-separated list of names of the hostgroups that we want the host to be part of. For example, in this case, we could have omitted the members directive in the hostgroup definition and specified a hostgroups directive, which has the value linux-servers, in the localhost host definition.

At this point, we have a localhost host and a linux-servers hostgroup, and localhost is a member of linux-servers. This is illustrated in the following figure:

Going further into localhost.cfg, we have a bunch of service object definitions that follow. Each of these definitions indicate the service on a localhost that we want to monitor with the host_name directive.

define service { use local-service host_name localhost service_description PING check_command check_ping!100.0,20%!500.0,60% }

This is one of the service definitions. The object defines a PING service check that monitors the reachability. The host_name directive specifies the host that this service check should be associated with, which in this case is localhost. Again, the value of the host_name directive here should reflect the value of the host_name directive defined in the host object definition. So, we have a PING service check defined for a localhost, which is illustrated by following figure:

There are several such service definitions that are placed on a localhost. Each service has a check_command directive that specifies the command for monitoring that service.

Note that the exclamation marks in the check_command values are the command argument separators. So, cmd!foo!bar indicates that the command is cmd with foo as its first argument and bar as the second.

It is important to remember that the check_ping part in check_command in the preceding example does not mean the check_ping executable that is in /usr/lib64/nagios/plugins/check_ping for most installations; it refers to the Icinga object of type command. In our setup, all command object definitions are inside commands.cfg.

The commands.cfg file has the command object definition for check_ping.

define command { command_name check_ping command_line $USER1$/check_ping -H $HOSTADDRESS$ -w $ARG1$ -c $ARG2$ -p 5 }

The check_command value in the PING service definition refers to the preceding command object, which indicates the exact command to be executed for performing the service check. $USER1$ is a user-defined Icinga macro. Macros in Icinga are like variables that can be used in various object definitions to wrap data inside these variables. Some macros are predefined, while some are user defined. These user macros are usually defined in /etc/icinga/resources.cfg:


So replace the $USER1$ macro with its value, and execute:

$ value/of/USER1/check_ping --help

This command will print the usual usage string with all the command-line options available. $ARG1$ and $ARG2$ in the command definition are macros referring to the arguments passed in the check_command value in the service definition, which are 100.0,20% and 500.0,60% respectively for the PING service definition. We will come to this later. As noted earlier, the status of the service is determined by the exit code of the command that is specified in the command_line directive in command definition.

We have many such service definitions for a localhost in localhost.cfg, such as Root Partition (monitors disk space), Total Processes, Current Load, HTTP, along with command definitions in commands.cfg for check_commands of each of these service definitions.

So, we have a host definition for localhost, a hostgroup definition linux-servers having localhost as its member, several service check definitions for localhost with check commands, and the command definitions specifying the exact command with arguments to execute for the checks. This is illustrated with the example Ping check in the following figure:

This completes the basic understanding of how our localhost monitoring is built up from plain-text configuration.


We would, as is the point of having monitoring systems, like to get alerted when something actually goes down. We don't want to keep monitoring the Icinga web interface screen, waiting for something to go down. Icinga provides a very generic and flexible way of sending out alerts. We can have any alerting script triggered when something goes wrong, which in turn may run commands for sending e-mails, SMS, Jabber messages, Twitter tweets, or practically anything that can be done from within a script. The default localhost monitoring setup has an e-mail alerting configuration.

The way these notifications work is that we define contact objects where we give the contact name, e-mail addresses, pager numbers, and other necessary details. These contact names are specified in the host/service templates or the objects themselves. So, when Icinga detects that a host/service has gone down, it will use this contact object to send contact details to the alerting script. The contact object definition also has the host_notification_commands and service_notification_commands directives. These directives specify the command objects that should be used to send out the notifications for that particular contact. The former directive is used when the host goes down, and the latter is used when a service goes down. The respective command objects are then looked up and the value of their command_line directive is executed. This command object is the same as the one we looked at previously for executing checks. The same command object type is used to also define notification commands.

We can also define contact groups and specify them in the host/service object definitions to alert a bunch of contacts at the same time. We can also give a comma-separated list of contact names instead of a contact group.

Let's have a look at our current setup for notification configuration. The host/service template objects have the admin contact group specified, whose definition is in contacts.cfg:

define contactgroup { contactgroup_name admins alias Icinga Administrators members icingaadmin }

The group has the icingaadmin member contact, which is again defined in the same file:

define contact { contact_name icingaadmin use generic-contact alias Icinga Admin email your@email.com }

The contacts.cfg file has your e-mail address. The contact object inherits the generic-contact template contact object.

define contact{ name generic-contact service_notification_period 24x7 host_notification_period 24x7 service_notification_options w,u,c,r,f,s host_notification_options d,u,r,f,s service_notification_commands notify-service-by-email host_notification_commands notify-host-by-email register 0 }

This template object has the host_notification_commands and service_notification_commands directives defined as notify-host-by-email and notify-service-by-email respectively. These are commands similar to what we use in service definitions. These commands are defined in commands.cfg:

define command { command_name notify-host-by-email command_line /usr/bin/printf "%b" "***** Icinga *****\n\nNotification Type: $NOTIFICATIONTYPE$\nHost: $HOSTNAME$\nState: $HOSTSTATE$\nAddress: $HOSTADDRESS$\nInfo: $HOSTOUTPUT$\n\nDate/Time: $LONGDATETIME$\n" | /bin/mail -s "** $NOTIFICATIONTYPE$ Host Alert: $HOSTNAME$ is $HOSTSTATE$ **" $CONTACTEMAIL$ } define command { command_name notify-service-by-email command_line /usr/bin/printf "%b" "***** Icinga *****\n\nNotification Type: $NOTIFICATIONTYPE$\n\nService: $SERVICEDESC$\nHost: $HOSTALIAS$\nAddress: $HOSTADDRESS$\nState: $SERVICESTATE$\n\nDate/Time: $LONGDATETIME$\n\nAdditional Info:\n\n$SERVICEOUTPUT$\n" | /bin/mail -s "** $NOTIFICATIONTYPE$ Service Alert: $HOSTALIAS$/$SERVICEDESC$ is $SERVICESTATE$ **" $CONTACTEMAIL$ }

These commands are eventually executed to send out e-mail notifications to the supplied e-mail addresses. Notice that command_lines uses the /bin/mail command to send e-mails, which is why we need a working setup of a SMTP server. Similarly, we could use any command/script path to send out custom alerts, such as SMS and Jabber. We could also change the above e-mail command to change the content format to suit our requirements.

The following figure illustrates the contact and notification configuration:

The correlation between hosts/services and contacts/notification commands is shown below:


In this article, we analyzed our current configuration for the Icinga setup which monitors a localhost. We can replicate this to monitor a number of other servers using the desired service checks. We also looked at how the alerting configuration works to send out notifications when something goes down.

Resources for Article:

Further resources on this subject:

Books to Consider

comments powered by Disqus

An Introduction to 3D Printing

Explore the future of manufacturing and design  - read our guide to 3d printing for free