intermediate
Grafana
Build dashboards and alert views that combine metrics, logs, traces, annotations, and operational context.
Grafana is the common visualization layer for metrics, logs, and traces. Dashboards combine panels from Prometheus, Loki, Tempo, CloudWatch, Datadog, and more. Alerting can run in Grafana or upstream sources.
| Feature | Use | |---------|-----| | Dashboards | SLO views, RED panels, infra USE | | Variables | Filter by env, service, cluster | | Annotations | Deploy markers on graphs | | Alert rules | Notify when queries breach thresholds |
Design dashboards for incident triage: above-the-fold golden signals, links to logs/traces with matching labels, and ownership in folder names.
On interviews: dashboard-as-code, avoiding vanity graphs, and unified views vs tool sprawl.
Common pitfalls: hundreds of unused dashboards; no standard templates; alerts duplicated in Grafana and Prometheus.
The trade-off is flexible visualization versus dashboard debt without governance.
Checklist:
- Standardize golden-signal templates.
- Link panels to logs/traces.
- Annotate deploys and incidents.
- Review unused dashboards periodically.