Any working examples of using the `collect_scrape_jobs` behavior?

[madelaney]

Affected Puppet, Ruby, OS and module versions/distributions

• Puppet:

$ sudo /opt/puppetlabs/bin/puppet --version
6.4.0

• Ruby:

$ sudo /opt/puppetlabs/puppet/bin/ruby --version
ruby 2.5.3p105 (2018-10-18 revision 65156) [x86_64-linux]

• Distribution: ??
• Module version: 8.3.0

How to reproduce (e.g Puppet code you use

prometheus server

class profile::prometheus::server {
  include ::prometheus
  firewall {
    '8093 Allow prometheus connections':
      proto  => 'tcp',
      dport  => [9100],
      action => 'accept';
  }
}

prometheus::manage_prometheus_server: true
prometheus::scrape_configs: []
prometheus::collect_scrape_jobs:
  - job_name: prometheus-node-exporter

prometheus client

class profile::prometheus::client {
  include ::prometheus
}

shared yaml

prometheus::scrape_configs:
- job_name: prometheus
  scrape_interval: 10s
  scrape_timeout: 10s
  static_configs:
  - targets:
    - stats.alertinnovation.com:9090
    labels:
      alias: prometheus
- job_name: nodexporter
  scrape_interval: 10s
  scrape_timeout: 10s
  static_configs:
  - targets:
    - stats.alertinnovation.com:9100
    labels:
      alias: node
prometheus::version: 2.4.3
prometheus::node_exporter::version: 1.0.0-rc.0

What are you seeing

Catalog is applied without any errors but the prometheus server isn't collecting any node_exporter configs.

What behaviour did you expect instead

I expect to see the catalog applied but also the prometheus server collecting node_exporter configs.

Output log

Any additional information you'd like to impart

I think this is more of a usage question then a bug but I feel the examples do not really show how to exercise this collector behavior.

[tomashejatko]

Hello @madelaney, I have not found much documentation on this, but this is my supersimple configuration made today, that works for me :)

shared yaml (common.yaml for me) Hiera :

prometheus::node_exporter::export_scrape_job: true
prometheus::apache_exporter::export_scrape_job: true
prometheus::collect_scrape_jobs:
  - job_name: node
  - job_name: apache

Node definition for Prometheus server (I want to clean this up and move to Hiera later) :

  class { 'prometheus::server':
    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.'
              }
            },
          ],
        },
      ],
    },
    scrape_configs => [
      {
        'job_name'        => 'prometheus',
        'scrape_interval' => '10s',
        'scrape_timeout'  => '10s',
        'static_configs'  => [
          {
            'targets' => [ 'localhost:9090' ],
            'labels'  => {
              'alias' => 'Prometheus',
            }
          }
        ],
      },
    ],
  }

Then I just include prometheus::node_exporter and prometheus::apache_exporter where I need them.

Just basic example, hope it helps!

[genebean]

I am currently trying to wrap my head around this too

[tomashejatko]

I can provider bit of update here. In my global Hiera config (common.yaml) I have:

lookup_options:
  prometheus::blackbox_exporter::modules:
    merge:
      strategy: deep
      merge_hash_arrays: true
  prometheus::node_exporter::collectors_enable:
    merge:
      strategy: deep
      merge_hash_arrays: true
prometheus::storage_retention: 90d
prometheus::daemon::export_scrape_job: true
prometheus::node_exporter::export_scrape_job: true
prometheus::node_exporter::collectors_enable:
  - systemd
  - processes
prometheus::apache_exporter::export_scrape_job: true
prometheus::blackbox_exporter::export_scrape_job: true
prometheus::elasticsearch_exporter::export_scrape_job: true
prometheus::haproxy_exporter::export_scrape_job: true
prometheus::collect_scrape_jobs:
  - job_name: apache
  - job_name: elasticsearch
  - job_name: haproxy
  - job_name: mysql
  - job_name: node
  - job_name: postgres
  - job_name: postfix
  - job_name: process
  - job_name: puppetdb
  - job_name: redis
prometheus::mysqld_exporter::export_scrape_job: true
prometheus::mysqld_exporter::cnf_socket: /var/run/mysqld/mysqld.sock
prometheus::mysqld_exporter::cnf_user: root
prometheus::node_exporter::collectors_enable:
  - mountstats
  - ntp
  - processes
  - systemd
prometheus::postgres_exporter::export_scrape_job: true
prometheus::postfix_exporter::export_scrape_job: true
prometheus::process_exporter::export_scrape_job: true
prometheus::puppetdb_exporter::export_scrape_job: true
prometheus::redis_exporter::export_scrape_job: true

Prometheus server:

classes:
  - grafana
  - prometheus::blackbox_exporter
  - prometheus::node_exporter

grafana::provisioning_datasources:
  apiVersion: 1
  datasources:
    - name: 'Prometheus'
      type: 'prometheus'
      access: 'proxy'
      url: 'http://localhost:9090'
      isDefault: true
grafana::provisioning_dashboards:
  apiVersion: 1
  providers:
    - name: 'default'
      orgId: 1
      folder: ''
      type: file
      disableDeletion: true
      options:
        path: '/var/lib/grafana/dashboards'
        puppetsource: 'puppet:///modules/arteal/grafana_dash'

prometheus::manage_prometheus_server: true
prometheus::postfix_exporter::manage_user: false
prometheus::postfix_exporter::user: postfix

Then host with Redis and Apache looks like:

classes:
  - apache
  - prometheus::apache_exporter
  - prometheus::redis_exporter
  - redis

And I use default node manifest like:

node "default" {

	hiera_include(classes)

}

This works perfectly Let me know if I can help somehow else

[genebean]

Part of what I am trying to figure out is how to define in a webserver's profile that I want a blackbox_exporter that lives on the promtheus sever node to test it. The idea is to have the backbox_exporter remotely check the webserver. This all works fine if I put static configs in the hiera of the prometheus server, but I want to put configs with the profile of the thing that needs to be checked