Skip to content

WIP: Monitoring - First Iteration

Administrator requested to merge start-monitoring into master

Related with the Trello card: https://trello.com/c/gWnVcNeE

Context

We decide the last week start to monitoring the OC servers with our new stack of monitoring (Grafana). The casualty, this stack is the ones recommended by the OC team: https://opencell.assembla.com/spaces/meveo-enterprise/wiki/Monitoring_and_Alerting

image

All these pieces can be installed in a Docker container, isolated, but this architecture expects the mount of volumes with all the data needed by the metric exporters (https://github.com/google/cadvisor/blob/master/docs/running.md#with-docker). Also we need a firewall to block the exposed ports.

By these reasons and because the exporters are small binaries that can be run without any dependencies installation, we decide to install all these pieces in the host machine. This is the first approximation.

TODO LIST

  • Review the pieces of the stack
  • Add Node Exporter
  • Add Prometheus
  • Add CAdvisor

By now, only expose the metric with NodeExporter to avoid security issues.

Grafana with Prometheus dataset and Node Exporter

Edited by Administrator

Merge request reports