[kube-prometheus-stack] kubernetes-mixin dashboards links do not work anymore

[FlorisFeddema]

Describe the bug a clear and concise description of what the bug is.

When using the kubernetes-mixin dashboads the links to more in-depth dashboard do not work.
For example if I open the Kubernetes / Compute Resources / Workload dashboard and click on one of the rows in the CPU Quota table it links to another dashboard. However the url of this dashboard does not seem to be working.

These are the parameters I end up with:

What's your helm version?

v3.10.3

What's your kubectl version?

v1.28.2

Which chart?

kube-prometheus-stack

What's the chart version?

56.6.2

What happened?

Went to the Kubernetes / Compute Resources / Workload dashboard and clicked on one of the rows in the CPU Quota table. The newly opened page does not have the correct paramters.

What you expected to happen?

The newly opened page to use the parameters of the tablerow to fill in the next dashboard parameters.

How to reproduce it?

Deploy kube prometheus stack with grafana and the dashboards.
Navigate to the Kubernetes / Compute Resources / Workload dashboard and click on one of the rows in the CPU Quota table.

Enter the changed values of values.yaml?

NONE

Enter the command that you execute and failing/misfunctioning.

The install worked.

Anything else we need to know?

No response

[fctb]

Click on pod name (drill down) forward not the pod name. The url contains: &var-pod=${__value.text} instead of the pod name.
Likly came from here: 4087
I use grafana 10.3.1

Could it be that new $vars don't work with old panels, which are deprecated baut here still used?
Or do we miss an dashboard update to switch to panels?

[FlorisFeddema]

I tried some things and I could not get the links working with the new vars indeed.

The links result in the following data fields:

[jkroepke]

Thats strange. I implement this, because the old table panel wont work with the old variables.

However I have multiple grafana instances, where is issues does not appear. Currently, I'm at 10.3.3, but it worked with 10.2.x.

For any reasons. Grafana always to an automatic migration of the table panel. Seems like this is not the case on your site.

Could you please generate an json export?

[FlorisFeddema]

Here is the export of our dashboard. We have about 30 environments where we run Grafana and they all have the same problem.
I have also deleted the pvc and recreated the grafana instance.
Kubernetes _ Compute Resources _ Workload-1708936097167.json

[jkroepke]

If you are try to edit the table panal, does Grafana offer the migration?

[FlorisFeddema]

I don't see any option to do the migration in Grafana. We run version 10.3.3

This is what we get when we edit the panel:

[jkroepke]

Ah. Do you have an test instance, where you could set on grafana.ini:

[security]
angular_support_enabled = false

Ref: https://grafana.com/docs/grafana/latest/setup-grafana/configure-grafana/#angular_support_enabled

Looking at source, if this is set, it will auto-migrate the old table:

https://github.com/grafana/grafana/blob/648abdbd0ea4451a102601dec874c770d635eb12/public/app/features/dashboard/state/DashboardModel.ts#L172-L174

This explain it to me and we may need an values toggle here.

[FlorisFeddema]

I have deployed it on one of our test instances and after disabling the angular support flag in the grafana config it automatically migrated the tables to the new one!

[jkroepke]

I implement a toggle so it will work for both scenario now, once the PR is merged.

[pierreozoux]

@jkroepke we think we have the same issue on the node dashboard.

we attached a PR to this bug report:
#4828

Did you modify upstream https://github.com/kubernetes-monitoring/kubernetes-mixin repo, or just here?
We'd like to submit a proper PR, but are a bit lost to be honest :) If you could this would be nice!

Thanks for your help!