Multicluster enhancements

[trent-s]

Consider how to enable multicluster support:

The end goal is to provide a way to useful ways to view energy/CO2 data for multiple clusters.
Possible deliverables include:

• A mechanism/dashboard (running anywhere?) that can query and display energy/CO2 for (selectable) labelgroups on (selecatble) clusters.
  • Seems that that CPU/CO2 data can be obtained and displayed if clusters and authentication (and certain configuration information) is provided.

Questions
- Can/should we ease deployment on multiple clusters?
- How exactly to gather and graph data for multiple clusters.... (just use config info and URL?)
- Could be treated as a separate project that depends on SusQL deployment.
- Do we need cluster identifier in data? maybe no, but seems like a good idea.
- Cache/store gathered data?

[trent-s]

Hopefully getting access to SusQL data should be as simple as accessing Thanos as follows (assuming logged into specific cluster):

export thanosroute=$(oc get routes -n openshift-monitoring thanos-querier -o jsonpath='{.status.ingress[0].host}')
curl -s  -k -H "Authorization: Bearer $(oc whoami -t)" https://${thanosroute}/api/v1/query?query=susql_total_energy_joules | jq

[mamy-CS]

Yes, I was told that Thanos does multiple cluster aggregation from multiple Prometheus instances in different clusters for observability. Explore this.