Grafana
Dashboards, panels, alerts, and data sources
Grafana
Grafana is a visualization platform for monitoring data. It connects to data sources (Prometheus, Loki, PostgreSQL, etc.) and displays metrics as beautiful, interactive dashboards.
Key Concepts
- Dashboard — collection of panels displaying related metrics
- Panel — individual visualization (graph, gauge, table, heatmap, stat)
- Data Source — backend to query (Prometheus, InfluxDB, MySQL, Loki, Elasticsearch)
- Variables — template variables for dynamic dashboards ($namespace, $pod)
- Alerts — threshold-based notifications sent to contact points
Common Dashboard Patterns
- RED Method (services): Rate, Errors, Duration — for request-driven services
- USE Method (resources): Utilization, Saturation, Errors — for infrastructure
- The Four Golden Signals (Google SRE): Latency, Traffic, Errors, Saturation
💬 RED vs USE method — when to use which?
RED Method: for services/endpoints — how many requests, how many errors, how long they take. USE Method: for infrastructure resources (CPU, memory, disk, network) — how utilized, how saturated, any errors. Use both together for complete observability.