puppet-prometheus

Compatibility

Prometheus Version	Recommended Puppet Module Version
>= 0.16.2	latest

node_exporter >= 0.15.0 consul_exporter >= 0.3.0

Background

This module automates the install and configuration of Prometheus monitoring tool: Prometheus web site

What This Module Affects

• Installs the prometheus daemon, alertmanager or exporters(via url or package)
  • The package method was implemented, but currently there isn't any package for prometheus
• Optionally installs a user to run it under (per exporter)
• Installs a configuration file for prometheus daemon (/etc/prometheus/prometheus.yaml) or for alertmanager (/etc/prometheus/alert.rules)
• Manages the services via upstart, sysv, or systemd
• Optionally creates alert rules
• The following exporters are currently implemented: node_exporter, statsd_exporter, process_exporter, haproxy_exporter, mysqld_exporter, blackbox_exporter, consul_exporter

Usage

To set up a prometheus daemon: On the server (for prometheus version < 1.0.0):

class { '::prometheus':
  global_config  => { 'scrape_interval'=> '15s', 'evaluation_interval'=> '15s', 'external_labels'=> { 'monitor'=>'master'}},
  rule_files     => [ "/etc/prometheus/alert.rules" ],
  scrape_configs => [
     { 'job_name'=> 'prometheus',
       'scrape_interval'=> '10s',
       'scrape_timeout' => '10s',
       'target_groups'  => [
        { 'targets'     => [ 'localhost:9090' ],
            'labels'    => { 'alias'=> 'Prometheus'}
         }
      ]
    }
  ]
}

On the server (for prometheus version >= 1.0.0):

class { 'prometheus':
    version => '1.0.0',
    scrape_configs => [ {'job_name'=>'prometheus','scrape_interval'=> '30s','scrape_timeout'=>'30s','static_configs'=> [{'targets'=>['localhost:9090'], 'labels'=> { 'alias'=>'Prometheus'}}]}],
    extra_options => '-alertmanager.url http://localhost:9093 -web.console.templates=/opt/prometheus-1.0.0.linux-amd64/consoles -web.console.libraries=/opt/prometheus-1.0.0.linux-amd64/console_libraries',
    localstorage => '/prometheus/prometheus',
}

or simply:

include ::prometheus

To add alert rules, add the following to the class prometheus in case you are using prometheus < 2.0:

    alerts => [{ 'name' => 'InstanceDown', 'condition' => 'up == 0', 'timeduration' => '5m', labels => [{ 'name' => 'severity', 'content' => 'page'}], 'annotations' => [{ 'name' => 'summary', content => 'Instance {{ $labels.instance }} down'}, {'name' => 'description', content => '{{ $labels.instance }} of job {{ $labels.job }} has been down for more than 5 minutes.' }]}]

or in hiera:

alertrules:
    -
        name: 'InstanceDown'
        condition:  'up == 0'
        timeduration: '5m'
        labels:
            -
                name: 'severity'
                content: 'critical'
        annotations:
            -
                name: 'summary'
                content: 'Instance {{ $labels.instance }} down'
            -
                name: 'description'
                content: '{{ $labels.instance }} of job {{ $labels.job }} has been down for more than 5 minutes.'

When using prometheus >= 2.0, we use the new yaml format (https://prometheus.io/docs/prometheus/2.0/migration/#recording-rules-and-alerts) configuration

alerts:
  groups:
    - name: alert.rules
      rules:
      - alert: 'InstanceDown'
        expr: 'up == 0'
        for: '5m'
        labels:
          'severity': 'page'
        annotations:
          'summary': 'Instance {{ $labels.instance }} down'
          'description': '{{ $labels.instance }} of job {{ $labels.job }} has been down for more than 5 minutes.'

On the monitored nodes:

   include prometheus::node_exporter

or:

class { 'prometheus::node_exporter':
    version => '0.12.0',
    collectors_disable => ['loadavg','mdadm' ],
    extra_options => '--collector.ntp.server ntp1.orange.intra',
}

For more information regarding class parameters please take a look at class docstring.

Limitations/Known issues

In version 0.1.14 of this module the alertmanager was configured to run as the service alert_manager. This has been changed in version 0.2.00 to be alertmanager.

Do not use version 1.0.0 of Prometheus: https://groups.google.com/forum/#!topic/prometheus-developers/vuSIxxUDff8 ; it does break the compatibility with thus module!

Even if the module has templates for several linux distributions, only RH family distributions were tested.