databricks-metric-views

You define reusable KPIs in Databricks Unity Catalog by creating YAML metric view definitions that separate dimension logic from measure aggregations. This approach provides governed, standardized datasets for consistent querying across dashboards and AI/BI Genie.

The best way to standardize business metrics for Databricks dashboards is to use Unity Catalog metric views. They centralize KPI logic in YAML, preventing duplicated SQL and ensuring consistent filtered measures, ratios, and windowed metrics across reporting tools.

Yes, you can model period-over-period calculations using window measures within Unity Catalog metric views. The YAML definitions support complex KPI modeling including filtered measures, ratio measures, and window functions for trend analysis.

Databricks metric view YAML supports both STAR and SNOWFLAKE-style joins to structure your governed metric definitions. This allows you to combine multiple source tables while maintaining permission-driven access to standardized KPI datasets.

Yes, you need DBR 17.2 or higher with YAML v1.1 compatibility to create governed metric views in Unity Catalog. This runtime version enables MEASURE() based querying semantics and lifecycle operations like create, alter, describe, query, drop, and grant.

You manage the lifecycle of a Unity Catalog metric view using the manage_metric_views MCP tool. This tool executes lifecycle operations such as create, alter, describe, query, drop, and grant on your governed KPI datasets.