Metrics Stack

This Docker Compose setup runs a self-contained observability stack with Prometheus, Grafana, Loki, Promtail, Tempo, and cAdvisor, wired for Traefik routing on a shared proxy network.

What this stack does

• Prometheus: time-series metrics storage and querying.
• Grafana: dashboards and visualization for metrics, logs, and traces.
• Loki: log storage backend.
• Promtail: log shipper to Loki (reads host and Docker logs).
• Tempo: trace storage backend.
• cAdvisor: container-level metrics exporter.

Note: node_exporter is expected to be running on the host machine and is not dockerized in this compose file.

Domains and environment

All user-specific values are set in .env and documented in .env.example. The DOMAIN variable defines the external hostname used by Traefik routing rules and service URLs.

Service exposure and hosting

• Grafana is exposed via Traefik at https://$DOMAIN/.
• Prometheus is exposed via Traefik at https://$DOMAIN/prometheus.
• Loki, Promtail, Tempo, and cAdvisor are not exposed publicly; they are only reachable within internal Docker networks.

Networks

• proxy (external): Shared Traefik network used to expose Grafana and Prometheus via HTTPS routers.
• pelican_nw (external): Existing external network used by Prometheus for upstream integrations.
• monitoring (bridge): Internal monitoring network for inter-service communication, with a fixed subnet and bridge name (br-monitoring).

Volumes

Named volumes persist state for:

• Prometheus (data-prometheus)
• Grafana (data-grafana)
• Loki (data-loki)
• Tempo (data-tempo)