feat: allow multiple kepler deployments using kepler-internal

sthaha commented 9 months ago

Important Changes:

allows setting deployment-namespace instead of hardcoded openshift-kepler-operator
modifies the default kepler deployment namespace to kepler-operator
Allows watching additional namespaces using watch-namepaces=foo,bar --watch-namespaces=baz (multiple list)
Adds status.exporter to indicate kepler deployment conditions

Testing the PR locally To give this a spin on kind cluster created using make fresh

operator-sdk run bundle quay.io/sthaha/kepler-operator-bundle:0.0.0-kepler-internal \
  --namespace=operators \
  --use-http

vprashar2929 commented 9 months ago

In case of multiple kepler deployments we have to figure out openshift dashboard's deployments and also revise our prometheus rules accordingly.

sthaha commented 9 months ago

In case of multiple kepler deployments we have to figure out openshift dashboard's deployments and also revise our prometheus rules accordingly.

Thats not a supported scenario to start with. I am thinking of adding a

spec:
  dashboards: 
      enabled: t|f
  exporter:
     ....

to KeplerInternal where dashboards.enabled defaults to false. I also think the prometheusrule may be giving wrong values since we are not filter out by service.

sthaha commented 9 months ago

@vprashar2929 I have added an openshift specific api now to enable / disable dashboards.

sustainable-computing-io / kepler-operator

feat: allow multiple kepler deployments using kepler-internal #309