Prometheus Metrics

Enable Prometheus metrics in the config:

monitoring:
  prometheus_enabled: true

Metrics are available at /metrics.

Available Metrics

Metric	Type	Description
`auto_ai_router_credential_rpm_current`	Gauge	Current RPM usage per credential
`auto_ai_router_credential_tpm_current`	Gauge	Current TPM usage per credential
`auto_ai_router_credential_banned`	Gauge	Ban status per credential (1 = banned)
`auto_ai_router_requests_total`	Counter	Total requests processed
`auto_ai_router_requests_duration_seconds`	Histogram	Request latency distribution

Proxy Credential Exclusion

Proxy credentials are not included in Prometheus metrics. Their statistics are available through the /health endpoint and are synchronized from the remote /health endpoint every 30 seconds.

Scrape Configuration

Example Prometheus scrape config:

scrape_configs:
  - job_name: 'auto-ai-router'
    scrape_interval: 15s
    static_configs:
      - targets: ['localhost:8080']