Q: How should I monitor my Docker BYOS M-Lab node, and what logging configuration is recommended?
A: Since M-Lab does not provide alerts or notifications for BYOS nodes, you must set up your own monitoring system.
Prometheus Metrics Endpoints
Deploy a Prometheus instance to scrape metrics from these endpoints in your Docker Compose stack:
| Service | Port | Key Metric |
|---|---|---|
| ndt-server | 9990/metrics | ndt7_client_test_results_total |
| jostler | 9991/metrics | — |
| uuid-annotator | 9992/metrics | — |
| heartbeat | 9993/metrics | — |
| traceroute-caller | 9994/metrics | — |
| node_exporter | 9995/metrics | CPU, memory, disk usage |
Critical Logging Configuration
Docker’s default json-file logging driver does not rotate logs, and M-Lab containers generate significant log output that can fill your disk. Configure Docker to use the local logging driver by adding this to /etc/docker/daemon.json:
{
"log-driver": "local"
}Then restart Docker. The local driver includes automatic log rotation and size limits to prevent disk space issues over time.