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

We're currently in the process of upgrading the Prometheus-Stack to the latest version, but in our test environment, we noticed a few issues with the API-Server Monitoring.
In the Dashboard "Kubernetes / API server", one can see that several Panels show no data. While the "Write Availability (30d)" works fine, the following panels show "No Data":

• Availability (30d) > 99.000%
• ErrorBudget (30d) > 99.000%
• Read Availability (30d)

In my tests, I noticed that for example the Read Availability works if I add the or vector(0) fallback to the sums:
Example Grafana Explore View (you need to have a Port-forward to Grafana on your local Port 3000, i.e. via kubectl port-forward svc/prometheus-grafana 3000:80)

What's your helm version?

version.BuildInfo{Version:"v3.17.1", GitCommit:"980d8ac1939e39138101364400756af2bdee1da5", GitTreeState:"clean", GoVersion:"go1.23.6"}

What's your kubectl version?

Client Version: v1.32.2 Kustomize Version: v5.5.0 Server Version: v1.32.2

Which chart?

kube-prometheus-stack

What's the chart version?

69.7.3

What happened?

After upgrading to the latest version of kube-prometheus-stack (69.7.3), the API-Server Availability Dashboard is broken. Several Panels show "No Data" when they should show an availability of up to 100%.

What you expected to happen?

The Panels should show an availability of 100% or less.

How to reproduce it?

1. Install the kube-prometheus-stack via Helm
2. Open Grafana:
   1. kubectl port-forward svc/prometheus-grafana 3000:80
   2. Open with Browser: http://localhost:3000
   3. Log in using default credentials (admin - prom-operator)
   4. Open the Dashboard "Kubernetes / API server"
   5. Wait a few minutes for Prometheus to gather data and calculate the rules

Enter the changed values of values.yaml?

NONE

Enter the command that you execute and failing/misfunctioning.

The broken Panels use the Prometheus-Recording-Rules for apiserver_request:availability30d with the verbs read and all.

Anything else we need to know?

Background: After upgrading from kube-prometheus-stack from 62.6.0 to 69.7.2, we've noticed alerts about alertname: KubeAPIErrorBudgetBurn. Checking out the API-Server Dashboard revealed that the availability Panels showed more than 100% (i.e. 5000% and growing).
As the Upgrade to 68 has updated the behaviour of different API-Server metrics, we've used the Admin-API to delete older Metrics. But then some panels could not be shown anymore.
The same issue can be reproduced in a freshly installed Installation of the Kube-Prometheus-Stack.

Edit: Improve comprehensibility of Link "Example Grafana Explore View"