Skip to content

Introduction to the Freshness Check

The Freshness check asserts that a table or file was updated recently. It reads when the container was last modified, measures how long ago that was, and reports a Shape Anomaly when the age exceeds the Maximum Age you set. It answers when data last arrived. Its counterpart, Volumetric, answers how much data arrived.

This section is the complete guide to the check. The Deep Dive covers the definition, properties, and evaluation semantics, explains where the last modification time comes from and how anomalies are reported, walks through production examples, and collects best practices. The How-tos are step-by-step tutorials for creating, editing, and deleting a check. The API page documents the payload for programmatic use, and the FAQ answers the most frequent questions.

Next Steps

  • How It Works

    The complete reference: definition, when the check is evaluated, where the last modification time comes from, the Maximum Age, and AI Managed Freshness checks.

    How It Works

  • Anomaly Reporting

    What the Shape Anomaly reports, how the age is measured, and why there is nothing to report per row.

    Anomaly Reporting

  • Examples

    Three production scenarios: an extract that stops arriving, a daily load that runs late, and a view measured by its timestamp field.

    Examples

  • Best Practices

    Guidelines for scheduling the Scan, setting the Maximum Age, giving views an incremental field, and working with AI Managed checks.

    Best Practices

  • Permissions

    The team permission each action needs: view, create, edit, archive, restore, and delete.

    Permissions

  • Create a Check

    Step-by-step tutorial for creating a Freshness check on a table or file.

    Create a Check

  • Edit a Check

    Step-by-step tutorial for changing the Maximum Age of a check.

    Edit a Check

  • Delete a Check

    Step-by-step tutorial for archiving a check and deleting it permanently.

    Delete a Check

  • Troubleshooting

    Common problems: no anomaly despite stale data, a check that is never asserted, and a check that passes while the pipeline is broken.

    Troubleshooting

  • API

    Payload shape and field notes for creating a Freshness check programmatically.

    API

  • FAQ

    Short answers to questions about when the check runs, where its time comes from, and AI Managed Freshness checks.

    FAQ