Look into the shared grafana dashboards

[moorepants]

Just saw this announcement:

https://discourse.jupyter.org/t/announce-grafana-dashboards-for-jupyterhub-on-kubernetes/7320

Maybe we should adopt and contribute to these (if useful). Someone should evaluate if these are useful to us and if so, we can adopt.

[TimStewartJ]

I've tried adding these dashboards to our cluster. The installation process itself went fairly smoothly, though half of the dashboards and their stats don't show up and thus aren't helpful. I'm not sure why they might be like this but I suppose it is a different in how our Prometheus is set up vs. how the dashboards expect it to be.

[moorepants]

We talked about this at the meeting and the dashboards look like they add a new view into the pod level. So let's get them working. There are some issues accessing some of the data sources. Some of the issues is the capitalization of "prometheous".

[TimStewartJ]

I took some time to explore this again as its a little related to #255. One thing I noticed is that the kube_pod_labels query doesn't seem to work as intended. It doesn't properly return metrics with the requested labels. This is definitely something that needs to be looked into.