Observe · Visualisation
Grafana
Grafana visualises data from Prometheus, Loki, Tempo and many other sources, with dashboards as code, unified alerting and role-based access.
On this page
What it does
Queries data sources, renders dashboards and manages alert rules and notification policies in one UI.
Where it fits
Observe stage: the shared view for engineers and on-call.
Quick start
helm install grafana grafana/grafana -n monitoring --set adminPassword="$(openssl rand -base64 24)"
# Provision dashboards from Git: mount JSON under /var/lib/grafana/dashboards with a provider configCommon pitfalls
- Dashboards edited in the UI drift from Git; provision them and disable UI edits for shared folders.
- Anonymous access left enabled on internal instances exposes internal metrics.
On this site
- Kubernetes Observability with Prometheus, Grafana and Loki: the chart’s dashboards, a provisioned workload-triage dashboard with metrics and logs on one screen, and a real OOM restart loop diagnosed on it.
Articles about Grafana
Observability intermediate
Kubernetes Observability with Prometheus, Grafana and Loki
kube-prometheus-stack, Loki in single-binary mode and Alloy on a kind cluster; the PromQL and LogQL that answer operational questions; low-cardinality log labels; a real OOM restart loop diagnosed end to end.