Prometheus & Grafana setup

The exporter is just an HTTP endpoint; Prometheus does the scraping and Grafana does the visualization.

Prometheus

Add a scrape_configs entry pointing at the exporter:

global:
  scrape_interval: 15s

scrape_configs:
  - job_name: "librechat-exporter"
    static_configs:
      - targets: ["exporter:9087"]

If the exporter has auth enabled, add the matching block (authorization, basic_auth, or oauth2).

scrape_interval: 15s is the default Prometheus interval — you do not need to match it to the exporter's REFRESH_INTERVAL or ADVANCED_REFRESH_INTERVAL. The exporter caches the last-computed values; Prometheus just reads whatever is current at scrape time. Setting a Prometheus scrape interval much shorter than 15 s is wasteful since values don't update faster than the exporter's own loops.

Grafana

Two paths:

Import the bundled dashboard

The repository ships a sample librechat-exporter-dashboard.json. In Grafana:

1. Dashboards → New → Import.
2. Upload librechat-exporter-dashboard.json or paste its contents.
3. Pick your Prometheus data source.
4. Save.

See the Grafana dashboard page for screenshots and what's inside.

Build your own

Every metric is documented on the Metrics reference. Use those names directly in Grafana panel queries — e.g.:

# Cost in the last 24 hours
librechat_transaction_cost_24h

# Messages per second over a 5-minute window
rate(librechat_message_count[5m])

# Adoption rate trend
librechat_adoption_rate_30d