feat: Add pyroscope-monitoring helm chart

[simonswine]

This contains a fully fledged observability deployment for Pyroscope OSS monitoring

Over time I would like to consolidate our internal dashboards and alerts into this helm chart. This also allows us to test/debug Pyroscope OSS in a local kind deployment.

In order to test this setup in your local kind:

$ make deploy-micro-services deploy-monitoring
# Once everything is up
$ kubectl port-forward deployment/pyroscope-monitoring 3000:3000
# Open grafana on http://localhost:3000

[simonswine]

@manab-pr please have a look, this PR also adds the dashboards/rules used into a subfolder: https://github.com/grafana/pyroscope/tree/7227037426b3f6bc4acf77df7c287e8026ce96ed/operations/monitoring

[manab-pr]

Thanks for the context, @simonswine! I've reviewed #4525 and understand the approach now.

The pyroscope-monitoring Helm chart is great for Kubernetes users who want a complete monitoring stack. However, I want to make sure we're still addressing
the original issue #3624.
The users in that issue were asking for pre-compiled dashboard files because they:

• Don't necessarily use Kubernetes/Helm
• Have existing Grafana instances and just want to import dashboards
• Can't/won't install jsonnet/jb to compile the mixin

For non-helm users: Will there be a way to extract/download the dashboard JSON files from the helm chart for manual import into Grafana? Or would it make sense to keep a small set of pre-compiled dashboards specifically for users who need standalone files, while the helm chart serves K8s users?

[simonswine]

@manab-pr Yes if you look here, you will find the rules and dashboards: https://github.com/grafana/pyroscope/tree/20251015_helm-chart-monitoring/operations/monitoring

CI will make sure that those files are in sync, when someone changes the helm chart.

Obviously when you just use dashboards/rules, no one guarantees that you run kube-state-metrics/cadvisor/node-exporter. Like we do through the helm chart.

[aleks-p]

Looks good to me. I gave it a try locally as well, with success. I just added one question about the deployment in the new chart.

[aleks-p]

This embeds and runs Prometheus, Loki, Tempo and Pyroscope binaries. Do we want a way to configure and prevent this, or are we ok with running Pyroscope and other applications inside the deployment for monitoring Pyroscope? :)

[simonswine]

This is exactly why I have chosen it, I don't want to maintain that gang of 4. I don't expect anyone to really do that to a production grade with this, will mention all of that in the README

[marcsanmi]

LGTM! needed to clean my laptop memory, but run like a charm 😁

[marcsanmi]

Nit: Is it worth adding a section listing the dashboards that will be created automatically? Maybe even integrate it with helm docs by adding it as an annotation in values.yaml ?