genlayer_global_monitor

Global monitoring stack with Prometheus (as a remote_write receiver) and Grafana. Simple local ports, no proxy, no TLS.

What you get:

• A remote_write endpoint at http://<host>:<PROM_PUBLIC_PORT>/api/v1/write
• Grafana served at http://<host>:<GRAFANA_PORT> pre-provisioned with the Prometheus datasource

Quick start

1. Prereqs

• Docker and Docker Compose

2. Configure

• Copy .env.example to .env and set values:
  • PROM_PUBLIC_PORT (default 9090)
  • GRAFANA_PORT and Grafana admin creds
  • Adjust PROM_RETENTION if needed

3. Launch

docker compose up -d

Configure senders (remote_write)

On each Prometheus (or Prometheus Agent) that should ship metrics to this global instance, add something like the following to its config:

global:
  external_labels:
    source: node-1            # unique per sender
    chain: mainnet            # e.g., network/tenant/project

remote_write:
  - url: http://<your-host>:<PROM_PUBLIC_PORT>/api/v1/write
    queue_config:
      capacity: 20000
      max_shards: 16
      max_samples_per_send: 10000
      max_backoff: 5m

Important: Ensure each sender has distinguishing external_labels to avoid series collisions and to make queries useful (e.g., source, chain, project).

Security notes

• This setup is intentionally minimal and uses plain HTTP.
• If exposing to the internet, put it behind your own reverse proxy or LB that handles TLS and any auth/rate limiting you need.
• Size disk and adjust PROM_RETENTION according to expected metric volume and cardinality.

File layout

• docker-compose.yml — services: Prometheus, Grafana
• prometheus/prometheus.yml — minimal config; receiver is enabled by flag
• grafana/provisioning/datasources/datasource.yml — pre-provisioned Prometheus datasource
• .env.example — copy to .env and fill