Skip to main content

Integrating with Beak AI Gateway

Beak AI Observability can be integrated with Beak AI Gateway to provide unified tracing, monitoring, and evaluation of all LLM calls across providers.

Why Integrate with Gateway?​

  • Unified Tracing: Track LLM calls from Studio, Gateway, and external apps in one place.
  • Cross-Provider Analytics: Compare performance, latency, and cost across LLM providers.
  • Centralized Monitoring: Detect issues, bottlenecks, and anomalies in real time.

How to Integrate​

  1. Enable Observability in Gateway

    • In the Gateway configuration, enable observability and specify the Observability endpoint.

  2. Configure API Keys and Permissions

    • Ensure Gateway has the necessary credentials to send trace data to Observability.

  3. Map Projects and Endpoints

    • Link Gateway projects to Observability projects for consistent trace organization.

  4. Test Integration

    • Make a test LLM call via Gateway and verify the trace appears in Observability.

Example Configuration​

observability:
  enabled: true
  endpoint: https://observability.beakops.com/api/traces
  api_key: YOUR_OBSERVABILITY_API_KEY

Best Practices​

  • Regularly review cross-provider analytics to optimize routing and cost.
  • Set up alerts for error rates or latency spikes.
  • Use Observability dashboards to monitor usage and compliance.

For setting up a new project, see Setting Up a New Project. For tracing details, see Understanding a Trace.