Honeycomb Anomaly Detection API
Anomaly Detection identifies Signals - patterns in your services that deviate from their historical baseline - and can notify Recipients when a Signal turns anomalous. Each Signal is scoped to a service in a dataset and monitors one of the following: - **`error_rate`**: The fraction of erroring requests, compared to a trained normal band. - **`presence`**: Whether the service is emitting spans at all. Signals progress through the following statuses: - **`onboarding`**: The Signal has been created but does not yet have enough data to train. - **`normal`**: The Signal is trained and monitoring for anomalies. - **`anomalous`**: The Signal has detected an anomaly. - **`off`**: The Signal has been disabled and will not fire. - **`ineligible`**: The Signal cannot train because its service is not sending enough traffic to establish a reliable baseline. This API allows you to list Signals, retrieve a Signal with its assigned Recipients, page through a Signal's historical anomalies, and update a Signal's enabled state, sensitivity, and Recipient assignments. ## Authorization The API key must have the **Manage Signals** permission. Learn more about [API keys here](https://docs.honeycomb.io/configure/environments/manage-api-keys/).