Grafana

Grafana is the visualisation tool for Prometheus/Thanos metrics and Loki logs.

Links

NGOT

• stg: https://grafana.obs-corp-stg.csfpriv.com/
• prd: https://grafana.obs-corp-prd.cloudservicesfactory.com/

Note

In NGOT, there is only one Grafana per production line.

In your PromQL/LogQL expressions, do not forget to add a label cluster or obs_client to limit your requests.

Caascad

• Zone Infra
  • prd: https://grafana-infra.infra-prd.caascad.com/
  • stg: https://grafana-infra.infra-stg.caascad.com/

• Zones Cloud

  • https://grafana-infra.ocb-$CLIENT.caascad.com/

  Note

  In Caascad, there is another Grafana for the client: https://grafana.ocb-$CLIENT.caascad.com/. It is dedicated to the client and Caascad admins should not connect to it except if they know what they are doing.

Deployment

Deploy the helm chart caascad-grafana with Trackbone.

Restarting Grafana

Grafana can be restarted at any time.

Impact :

• the visualisation tool is unavailable during the restart
• any runtime modifications will be lost. !!! note Runtime modifications are discouraged because they can be lost at any time. Modifications should be done in Kubernetes objects.

Grafana datasources and dashboards are loaded at runtime. If a change on an existing dasboard is ignored, try to delete the dashboard with kubectl delete <configmap>, notice that it disappear from Grafana, then redeploy it. For datasources, try the same method.