Monitoring with Tecton
Tecton provides tools to monitor your feature engineering pipelines and the data they produce. This page provides an overview of the monitoring capabilities offered by Tecton and links to more detailed documentation on how to use each capability.
Data Quality Metrics
Data Quality Metrics help detect feature data issues quickly by providing summary statistics about the data produced by your Feature View pipelines. These metrics enable you to validate that your feature pipelines are writing the expected values into your feature store. Data Quality Metrics are available for Batch and Stream Feature Views.
Data Quality Validation
Data Quality Validation helps detect feature data issues once a Feature View has been materialized. If validation results indicate that feature data failed to meet expectations during a materialization interval, an alert email will be sent. Data Quality Validation is available for Batch and Stream Feature Views.
Alerting allows Tecton to automatically generate and send alerts for issues detected with your feature engineering pipelines and data. Tecton can alert on feature freshness, materialization failures, and data quality issues.
Monitoring Materialization provides details on how to monitor the health and freshness of your feature data using the Tecton Web UI, SDK, and CLI.
Production SLOs outlines Tecton's Service Level Objectives (SLOs) for feature serving latency and reliability. The SLOs provide guarantees around the performance of feature data served from Tecton in production.
Feature Request Audit Logging
Feature Request Audit Logging allows you to see what requests are being sent to your production feature serving endpoint. Feature Request Audit Logging must be enabled by Tecton Support.