Skip to content

Release 9.2.32.1

This release adds model usage tracking for Databricks ai_query invocations—including a new model detail page with a Usage tab—and keeps the asset inventory in sync automatically when agents change through an approved workflow.

AI Governance

Model usage tracking for Databricks ai_query

Model usage tracking for Databricks ai_query

  • Added a model asset detail page with a Usage tab, patterned after the agent Usage tab: Unique Users, Total Sessions, Active Days, and Avg Tokens/Day, with a time-range selector.
  • The AI Assets Collector now captures usage for Databricks ai_query invocations (foundation models and serving endpoints) that were previously not reflected in MLflow usage tracking, so model assets show real usage instead of an empty value.
Automatic inventory sync for approved agent workflows

Automatic inventory sync for approved agent workflows

  • When an AI agent is created or deleted through an approved workflow, the asset inventory now updates automatically. Previously, changes were not reflected until a manual asset sync or scan was performed.