add mimir continous test dashboard

[QuantumEnigmaa]

Towards giantswarm/roadmap#3578 and related to https://github.com/giantswarm/shared-configs/pull/149 and giantswarm/prometheus-rules#1355

This PR adds a Mimir / Continous test dashboard to display the mimir-continous-test component's metrics and have an overview of the current general Health of mimir.

There's an issue though as when I deployed the continuous-test component on grizzly the "failed" metrics for both queries and writes were missing while they are defined and referred to upstream. So before merging this, we need to get those metrics back.

Otherwise, here is the dahboard's current look :

Checklist

• Update changelog in CHANGELOG.md in an end-user friendly language.

[TheoBrigitte]

How can we have a 0.06 as value for Total queries test and other ? I read that as the number of queries tested, did I read this wrong ? Maybe we need a panel to explain what those values are.

[hervenicol]

So before merging this, we need to get those metrics back.

Does it mean this PR should be set as draft for now, and is not ready for review yet?

[QuentinBisson]

Can you add the same tags as the other mimir dashboards?

[QuentinBisson]

We need the same as

dashboards/helm/dashboards/charts/private_dashboards_mz/dashboards/shared/private/mimir-compactor-resources.json

Line 17 in 7af76dc

"links": [

[QuentinBisson]

I deployed this on golem and the metrics are fine so this is not a draft anymore.

@TheoBrigitte what do you mean with your idea?

Currently used metrics are defined here https://grafana.com/docs/mimir/latest/manage/tools/mimir-continuous-test/#exported-metrics if that helps

[QuentinBisson]

I added missing tags so it's up for reviews

[hervenicol]

There's an issue though as when I deployed the continuous-test component on grizzly the "failed" metrics for both queries and writes were missing while they are defined and referred to upstream. So before merging this, we need to get those metrics back.

Did these metrics get fixed?
Maybe the screenshot should be updated then?

[QuentinBisson]

Here you go

[hervenicol]

Once you've set a rate it's not total anymore.

[hervenicol]

LGTM.
I guess it will live and evolve when we actually use it then!