Grafana Dashboards

The fastest path from data to dashboard.

Step 1: Run Grafana

docker run -d -p 3000:3000 grafana/grafana

Open localhost:3000; log in as admin/admin.

Step 2: Connect Prometheus

• Configuration → Data Sources → Add data source → Prometheus.
• URL: http://prometheus:9090 (or wherever yours runs).
• Save & Test should turn green.

Step 3: First panel

The first panel is the proof that data is flowing. Pick a metric you trust, write the simplest useful query, and confirm the line draws.

• Create. Dashboards, New, Add panel; pick the Prometheus data source from the dropdown.
• Query. rate(http_requests_total[5m]); rate over a 5-minute window is the canonical request-rate query.
• Title. Name it 'Request Rate'; future-you reads dashboards like dashboards, not like queries.
• Apply. Save the panel and the dashboard; bookmark the URL for the team channel.

Step 4: Variables and templating

Variables turn one dashboard into a template. Pick a label, parameterise the query, and one dashboard now serves every service in your stack.

• Define a variable. Dashboard settings, Variables, New; type 'Query'; data source Prometheus; query label_values(http_requests_total, service).
• Reference in panels. rate(http_requests_total{service="$service"}[5m]); the dropdown drives every panel.
• Multi-value. Allow multi-select to compare services side by side without authoring a second dashboard.
• Reuse. One templated dashboard beats fifty hand-rolled per-service dashboards; drift drops to zero.

Antipatterns

• Hardcoding service names. Variables defeat that.
• One dashboard per service. Templates over copy-paste.
• Default panel without thoughtful query. Misleading.

What to do this week

Three moves. (1) Run the tutorial end-to-end on your own laptop / sandbox. (2) Apply the pattern to one production workload. (3) Document the variations you needed; share with the team.