Puppet module for prometheus
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.github
examples
files
manifests
spec
templates
.fixtures.yml
.gitignore
.msync.yml
.pmtignore
.rspec
.rubocop.yml
.sync.yml
.travis.yml
.yardopts
Gemfile
LICENSE
README.md
Rakefile
metadata.json

README.md

puppet-prometheus

Puppet Forge Puppet Forge Puppet Forge

Compatibility

Prometheus Version Recommended Puppet Module Version
>= 0.16.2 latest

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

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:

    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.'

On the monitored nodes:

class { '::prometheus::node_exporter':
  collectors => ['diskstats','filesystem','loadavg','meminfo','netdev','stat,time']
}

or:

class { 'prometheus::node_exporter':
    version => '0.12.0',
    collectors => ['diskstats','filesystem','loadavg','meminfo','logind','netdev','netstat','stat','time','interrupts','ntp','tcpstat'],
    extra_options => '-collector.ntp.server ntp1.orange.intra',
}

or simply:

include ::prometheus::node_exporter

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.