Overview

Nagios is the selected alerting/monitoring platform that Savannah uses. Nagios configuration and custom plugins can be found in the administration Git repository.

Throughout the rest of the wiki page \$NAG_CONF means the directory /etc/nagios3/conf.d

Access

The Nagios web interface is available at http://nagios.savannah.gnu.org/. It is protected by Basic Auth. To gain access, add yourself to /etc/nagios3/htpasswd.users

htpasswd /etc/nagios3/htpasswd.users username

Alerting

Nagios will send email alerts to users configured in \$NAG_CONF/contacts.cfg. As well as individual users, the savannah-reports-private mailing list is configured as a contact.

Configuration

  • Nagios runs on mgt.savannah.gnu.org
  • All hosts run NRPE to do active load, disk, users and processor checks
  • /opt/administration/nagios/conf.d is symlinked /etc/nagios3/conf.d

Service group vs host group

Any back-end checks that are global across all systems should have their checks in the host group. _Services_ that Savannah provides (Various VCS systems, WWW including static content and dynamic, DNS etc...) should be configured as a service and use the "virtual" host name of the system (EG: git.savannah.gnu.org is a CNAME for vcs.savannah.gnu.org)

Adding a new host

Edit:
\$NAG_CONF/hosts.cfg

and add the new host.

Edit:
\$NAG_CONF/hostgroups.cfg

and add the host to the all host group.

Check the configuration:
/etc/init.d/nagios3 reload # It will do a configtest and abort the reload on failure

Your new host should appear in Nagios under the host groups

Adding a new Plugin

Plugins for Nagios are simply scripts/binaries that follow specific exit codes and messages to stdout.

Custom plugins are located in /opt/administration/nagios/plugins. When creating a new plugin, add it here. Plugins follow a common naming convention of check_{name_of_service}.

Plugins should return three exit status.

  • GREEN - The script should exit with a code of 0 and print "OK - extra text..." to stdout
  • YELLOW - The script should exit with a code of 1 and print "WARNING
    • extra text..." to stdout
  • RED - The script should exit with a code of 2 and print "CRITICAL - extra text..." to stdout

Once a new plugin has been created, it will need to be added as a command in Nagios. Custom commands should be added to \$NAG_CONF/custom_commands.cfg.

Once the new command has been created, a service can then be added that uses this check_command.

Adding a new Service

Services map a check command to a host/service group. Add services to \$NAG_CONF/services.cfg. Look in this file for examples of configuring services.