v0.122.0

Version 0.121.0 adds list and list_effective methods for workspace grants, CPU workload options for model serving, and new fields across jobs and ML experiments. Breaking changes mandate the role field for Postgres database specs and remove select secret fields and ML aggregation functions.

The SDK now supports Genie attachment visualizations, SQL conditions for job schedules and triggers, and telemetry configuration for model serving endpoints. A breaking change removes the name field from the IAM v2 User model in favor

Databricks Python SDK v0.119.0 adds cluster policy compliance controls to track and cancel pending enforcements, alongside support for AI runtime tasks in Jobs. The

This release adds CRUD methods for workspace Postgres Data APIs and introduces Dynamics 365 as a Unity Catalog connection type. It also adds serverless compute ID fields for pipelines, endpoint IDs for vector search indexes, and capacity reservation group fields for Azure compute.

This release fixes WorkspaceClient construction crashes on Spark Connect clusters by making dbutils lazy and introducing a remote runtime fallback. It also adds OIDC token caching for Workload Identity Federation and updates API schemas

models with new fields, including custom retention hours in Unity Catalog, clustering configurations for pipelines, and faceting support in Vector Search.