Munin as optional part of the stack

[fititnt]

• References
  • Main page: http://munin-monitoring.org/
  • Documentation: http://guide.munin-monitoring.org/en/latest/
  • Plugins:
  • Core:
    • Gallery http://gallery.munin-monitoring.org/
    • Source https://github.com/munin-monitoring/munin/tree/master/plugins
  • 3rd:
    • Gallery http://gallery.munin-monitoring.org/contrib/index.html
    • Source https://github.com/munin-monitoring/contrib/tree/master/plugins
  • Galera Cluster
    • https://github.com/kjellm/munin-mysql/blob/master/contrib/Munin/MySQL/Graph/Galera.pm
    • https://www.percona.com/forums/questions-discussions/percona-monitoring-plugins/48494-percona-monitoring-plugin-for-munin

---

I'm still looking for strategies to monitor clusters (maybe not necessary including alerting as main role). Munin maybe fill this gap and may have a good combo:

1) Out of the box already have some plugins 2) Allow implement both node + "master" on same node or choose one node to collect data and others nodes just give information 3) It's not resource hungry by default (run as cron each 5 min, can be reduced to 1 min, but may not sufficient for alerting) 4) Have graphics

The alternatives I was looking where Prometheus + Grafana (both that actually may be showcased as demo too). In fact Prometheus started as a need to replace Munin because reducing just to a 1 minute was not sufficient.

At this moment I'm just testing on private repositories. It may become later source to publish on this public repo.

---

Edits

1. Added repository source codes for core and 3rd-party

[fititnt]

Humm... We will need some work on automate installation from 3rd party plugins from https://github.com/munin-monitoring/contrib/tree/master/plugins.

Most ansible roles just do with core plugins. Ok that is not complicated add option to download the extra plugins, on some pre_tasks at the playbooks