Monitoring

Monitoring has been built with the opensource tool Grafana. The URL of the Grafana access is https://base_url:/grafana and the credentials must be provided by Quobis support. It shows different dashboards with the live metric/logs values and metric history. By default the following dashboards are included:

_images/grafana_1.png

In order to provide a first check of the system, the Kubernetes cluster monitoring dashboard where the resources of the cluster is shown. Below you can see a screenshot of a running system during five working days:

_images/grafana_2.png

This is another screenshot during a normal day, where we can see the correlation between connected users, concurrent conferences and the network bandwith:

_images/grafana_2bis.png

It also provides the alerting service, where the following alerts are included by default in the deployment:

  • VM CPU/RAM/DISK usage.

  • Internal traffic load (inside the cluster)

  • Pods and container CPU/RAM/DISK usage.

  • Database status

_images/grafana_3.png

Example of alerts fired in the system

Several alerts can be added in function on the needs of the deployment. The following image describes the configuration of an alert:

_images/grafana_4.png

Example of the alert configuration