DevOpsil
monitoringMedium

Fix: Grafana Dashboard Showing 'No Data'

Grafana panels display 'No data' despite metrics existing in Prometheus

!Symptoms

  • Grafana panels display 'No data' despite metrics existing in Prometheus
  • Some panels work, others show no data
  • Dashboard worked previously but stopped after a change
  • PromQL queries work in Prometheus UI but not in Grafana

?Root Causes

  • Data source URL incorrect or pointing to wrong Prometheus instance
  • Time range in Grafana doesn't match when metrics were scraped
  • Metric name changed after relabeling or Prometheus upgrade
  • Grafana variable values not matching actual label values
  • Prometheus scrape interval longer than Grafana min interval

#Diagnosis Steps

  1. 1Test the query directly in Grafana's Explore tab
  2. 2Check data source configuration: Settings → Data Sources → Prometheus → Test
  3. 3Verify the metric exists: run the PromQL query in Prometheus UI
  4. 4Check Grafana variables: are they resolving to correct values?
  5. 5Compare time ranges between Grafana panel and Prometheus data retention

>Fix

  1. 1Fix data source URL to point to the correct Prometheus endpoint
  2. 2Adjust the time range to match when data was collected
  3. 3Update metric names in queries if they were renamed
  4. 4Fix template variable queries to match current label values
  5. 5Set min interval to match Prometheus scrape interval

*Prevention

  • Use Grafana provisioning to version-control dashboard JSON and data sources
  • Set up alerts for data source health checks
  • Document metric name changes in changelogs
  • Use Grafana's built-in metric browser to discover available metrics
  • Test dashboards after Prometheus configuration changes

Related Error Messages

No dataData source is not configured properlyError reading PrometheusClient error: 400 Bad Request