Grafana Dashboard

[giz33]

Hello!
I'm trying to understand how to use the https://github.com/pyrra-dev/pyrra/tree/main/examples/grafana.
And it says that I need to use --generic-rules, but where??
I need to use this flag when I am deploying Pyrra? Or when I am deployng the SLO's manifests inside examples/kubernetes/manifests/slos? Or I need to edit one of these manifests to put the --generic-rules inside ?

Thanks in advance for any help.

[metalmatze]

You need to run two Pyrra components.
One with the api command and one with the filesystem or kubernetes command. The filesystem or kubernetes components need to be run with this --generic-rules flag.

Hopefully you could figure it out in the mean time.
I'll make sure to update the documentation there to make it more obvious.

[giz33]

Hello!Tks for the answer!
I am trying to use the helm chart to deploy it and use with Grafana: https://artifacthub.io/packages/helm/rlex/pyrra
So far so good I am able to install it and run a pyrra-kubernetes changing the parameter generic-rules to true inside values.yaml.

But it keeps showing nothing on my Grafana dashboards.

I suspect about two things:
1 - I need to pass some args to pyrra-kubernetes using extraKubernetesArgs to point to my correct PrometheusURL, as you can see on the logs its using localhost:9090, but I have prometheus running on prometheus-operated.monitoring.svc.cluster.local:9090.
But I don't know which args I can pass on this parameter extraKubernetesArgs on values.yaml to change this.

2 - As you can see, it starts a metrics server on port 8080, but I don't see any exposed service to make my prometheus scrap this.And again, maybe there is some parameter that I can pass using extraKubernetesArgs and I don't know which parameters I can use(I didn't fin any documentation about parameter that I can use on it).

Tks in advance for any help.

[giz33]

I was able to make it works!
After identify that the prometheus operator uses a label to connect to all the objects that makes part of the solution I see that the components created by Parry doesn't have the correct label on it.
After change the label on rules and the SLO's objects to match the label used by my prometheus created by kube-prom operator, it works!

Thanks!

[stoopman]

I was able to make it works! After identify that the prometheus operator uses a label to connect to all the objects that makes part of the solution I see that the components created by Parry doesn't have the correct label on it. After change the label on rules and the SLO's objects to match the label used by my prometheus created by kube-prom operator, it works!

Thanks!

@giz33 can you please elaborate on which label specific this is? I have kube-prometheus-stack helm chart installed which includes the prometheus operator but no generic rules metrics (such as pyrra_objective) are being created in prometheus

[giz33]

@stoopman I put the label "release: prometheus"