Description of problem

In 4.18, the KubeJobFailed alert page fails to display metric results, and instead displays an error from /api/prometheus-tenancy/api/v1/query_range?start=1733162830.27&end=1733166430.269&step=12&namespace=%23ALL_NS%23&query=kube_job_failed%7Bjob%3D%22kube-state-metrics%22%2Cnamespace%3D%7E%22%28openshift-.*%7Ckube-.*%7Cdefault%29%22%7D+%3E+0&timeout=60s:

An error occurred
conflicting label matcher: label matcher "namespace=~\"(openshift-.*|kube-.*|default)\"" conflicts with injected matcher "namespace=\"#ALL_NS#\""

Version-Release number of selected component

I've seen the bug in 4.18.0-ec.4.

I have tested 4.17.7, and that worked, without reproducing this bug.

How reproducible

Every time.

Steps to Reproduce

1. launch a ClusterBot cluster with launch 4.18.0-ec.4 gcp.
2. create a broken Job (e.g. via attached YAML).
3. find the pending alert page in the in-cluster web console

Actual results

The Description's conflicting label matcher message.

Expected results

Successfully evaluated metrics.