f. Run
Rebuild and start the full stack:
Verify that all services β including prometheus and grafana β are running:
Expected output should show all containers with status running. If prometheus fails to start, double-check that setup/prometheus/prometheus.yml exists and is correctly formatted.
To confirm metrics are being scraped, open the Prometheus UI, navigate to Status β Targets, and verify that gateway, auth, and account all show state UP.
Troubleshooting
| Symptom | Cause | Fix |
|---|---|---|
| β target DOWN, server returned HTTP status 404 | metrics_path β the service's base-path + /prometheus | make the two match exactly (Microservices) |
| β target DOWN, no such host or connection refused | the service is not running, or the target says localhost | start it; inside a container localhost is the container itself, so use the Docker hostnames (gateway:8080β¦) |
β prometheus does not start | setup/prometheus/prometheus.yml missing or invalid YAML | check the file and the $SETUP path in .env |
| β οΈ dashboard panels empty | the DS_PROMETHEUS variable does not point to the datasource | create it as a constant with value prometheus, the uid in datasources.yml (Grafana) |
β
only /β¦/actuator/prometheus answers | exposure.include: ['prometheus'] | expected: the other Actuator endpoints stay closed |
Done! The microservices are now fully observable. Navigate to Grafana, import a dashboard, and start monitoring request rates, JVM memory, response latencies, and error rates across the entire platform.