Grafana Dashboards

Reproducible steps to build a first dashboard from the Grant API Prometheus metrics. High signal: exact PromQL and panel type.

Prerequisites

• API running with METRICS_ENABLED=true (and APP_PORT=4000 or match observability/prometheus.yml).
• Prometheus and Grafana up: docker compose up -d prometheus grafana.
• See observability/README.md in the repo for the full runbook.

1. Add Prometheus data source

1. Open Grafana (e.g. http://localhost:3001). Log in (default admin / admin or value of GRAFANA_ADMIN_PASSWORD).
2. Go to Configuration → Data sources → Add data source.
3. Choose Prometheus.
4. URL: http://prometheus:9090 (Grafana and Prometheus on same Docker network). If Grafana runs on the host, use http://localhost:9090.
5. Save & Test.

2. Create a dashboard

1. Dashboards → New → New dashboard.
2. Add visualization (or Add → Visualization).

3. Panel 1 — Request rate

1. Data source: Prometheus.

2. Query: (PromQL)

   rate(http_requests_total[5m])

3. Visualization: Time series.

4. Panel title: Request rate.

5. Unit: optional reqps (requests per second) or leave default. Apply or Save.

4. Panel 2 — P95 latency by route

1. Add → Visualization (new panel).

2. Query:

   histogram_quantile(0.95, sum(rate(http_request_duration_seconds_bucket[5m])) by (le, route))

3. Visualization: Time series.

4. Panel title: P95 latency by route.

5. Unit: s (seconds). Legend: . Apply or Save.

5. Save dashboard

Save dashboard; give it a name (e.g. "Grant API"). You can add more panels (e.g. error rate, request count by status) using the same data source and http_requests_total / http_request_duration_seconds_* metrics.

---

Related:

• Observability overview — Ports and adapters, config
• Runbook: observability/README.md in the repo