[kots]: add monitoring graphs

[mrsimonemms]

Description

Adds default monitoring graphs to the KOTS Graphs section of the app config. This allows users to connect to their own Prometheus instance and for the Gitpod instance to give out simple graphs. The graphs are based upon the published graphs in our monitoring docs.

NB we have limited (documented) control over the Y axis, so we end up with negative time periods by default. My feeling is that this is a convenience inside the KOTS dashboard and that, ultimately, most people will use Grafana over this. However, this should remain as a convenience for those who want a limited monitoring solution over a full-blown Grafana instance.

How to test

Install a Prometheus instance to your cluster and set the

As a convenience, my monitoring repo is already set up - the URL is http://monitoring-prometheus-prometheus.gitpod.svc.cluster.local:9090

NAMESPACE=gitpod
GITPOD_NAMESPACE=gitpod

helm upgrade \
  --atomic \
  --cleanup-on-fail \
  --create-namespace \
  --install \
  --namespace="${NAMESPACE}" \
  --repo=https://helm.simonemms.com \
  --reset-values \
  --set gitpodNamespace="${GITPOD_NAMESPACE}" \
  --wait \
  monitoring \
  gitpod-monitoring

Release Notes

[kots]: add monitoring graphs

Documentation

Werft options:

• /werft with-preview

[mrsimonemms]

Awaiting response to question

[nandajavarma]

LGTM! 📈