Skip to main content

Monitoring

Track your model's performance and usage with detailed monitoring and analytics.

Request Details​

For every prediction request, Heimdall records the following details:

  • Endpoint: The API endpoint that was called
  • Inference Count: Number of predictions made in this request
  • Response Time: How long the prediction took to complete
  • Drift %: Model performance drift detection
  • Channel: Where the request came from (UI, API, etc.)
  • User Agent: Browser/client information

Production Monitoring​

For more detailed monitoring information, see the Production Monitoring page which covers comprehensive monitoring, performance tracking, and production best practices.