A label normalization cleanup lands and later Kubernetes workload dashboards lose hierarchy views even though the source targets remain up.
A Grafana dashboard for kube-state-metrics goes blank because the relabel config now drops owner_kind while the dashboard variables still join on it for higher-level workloads
Scrapes are healthy, yet workload panels are empty because the join key used by the dashboard was stripped out during relabel cleanup.
Scenario
What to check first
- Identify the primary failure signal in the The Metric Survived and the Label the Dashboard Needed Did Not scenario.
- Separate visible symptoms from the underlying technical dependency.
- Describe the safest recovery path and the follow-up prevention work.
Checking checklist
- Summarize the current impact and the last known change.
- Collect direct evidence from logs, runtime state, and configuration before changing anything.
- Separate immediate recovery from permanent prevention work.
Recovery and prevention
Inspect the final metric label set before changing the dashboard queries or scrape intervals.
Questions worth viewing together
Community-field Kubernetes problem inspired by Grafana and Prometheus community patterns where relabeling removed owner_kind and blanked workload dashbo... Dashboard emptiness can be a post-scrape label problem even when collection is perfectly healthy.
Teams often blame kube-state-metrics instability when the relabel stage removed the ownership label dashboards joined on.
Relabel cleanup should be regression-tested against dashboard joins and template variables, not only ingestion volume.
Similar cases seen in the field