ReleaseBytes
databricks Databricks Python SDK Releases · · 0.106.0

Databricks SDK for Python v0.106.0: New Features and Breaking Changes

sdkinfradatabricksengineer
breaking patch

Databricks SDK for Python version 0.106.0 introduces new workspace and account-level services, including temporary volume credentials, enhanced knowledge assistant capabilities, and disaster recovery features. It also adds numerous fields to existing services for data pipelines and Postgres management, alongside a breaking change in how connections and tools are handled in the supervisor agents service. This release affects developers using the Databricks SDK, particularly those interacting with workspace services or relying on the supervisor agents' connection fields.

  • →Change name field for databricks.sdk.service.supervisoragents.Connection to no longer be required
  • →Change description field for databricks.sdk.service.supervisoragents.SupervisorAgent to no longer be required
  • →Remove connection field for databricks.sdk.service.supervisoragents.Tool
  • →Detect Databricks CLI version at init time via databricks version, enabling version-gated flag support without additional subprocess calls
  • →Add w.temporary_volume_credentials workspace-level service
Breaking changes (3) ›
  • Change name field for databricks.sdk.service.supervisoragents.Connection to no longer be required
  • Change description field for databricks.sdk.service.supervisoragents.SupervisorAgent to no longer be required
  • Remove connection field for databricks.sdk.service.supervisoragents.Tool
Enhancements (22) ›
  • Add w.temporary_volume_credentials workspace-level service

  • Add get_permission_levels(), get_permissions(), set_permissions() and update_permissions() methods for w.knowledge_assistants workspace-level service

  • Add undelete_project() method for w.postgres workspace-level service
  • Add thumbnail_url field for databricks.sdk.service.apps.App
  • Add confidential_compute_type field for databricks.sdk.service.compute.GcpAttributes
  • Add jira_options, outlook_options and smartsheet_options fields for databricks.sdk.service.pipelines.ConnectorOptions
  • Add google_ads_config field for databricks.sdk.service.pipelines.SourceConfig
  • Add replace_existing field for databricks.sdk.service.postgres.CreateBranchRequest
  • Add replace_existing field for databricks.sdk.service.postgres.CreateEndpointRequest
  • Add purge field for databricks.sdk.service.postgres.DeleteProjectRequest
  • Add show_deleted field for databricks.sdk.service.postgres.ListProjectsRequest
  • Add delete_time and purge_time fields for databricks.sdk.service.postgres.Project
  • Add uc_connection field for databricks.sdk.service.supervisoragents.Tool
  • Change name field for databricks.sdk.service.supervisoragents.Connection to no longer be required
  • Add databricks.sdk.service.disasterrecovery package
  • Add a.disaster_recovery account-level service

  • Add create_example(), delete_example(), get_example(), list_examples() and update_example() methods for w.knowledge_assistants workspace-level service

  • Add delta_table_name field for databricks.sdk.service.ml.BackfillSource
  • Add confluence_options field for databricks.sdk.service.pipelines.ConnectorOptions
  • Add confluence enum value for databricks.sdk.service.catalog.ConnectionType
  • Add confluence enum value for databricks.sdk.service.pipelines.IngestionSourceType
  • Change description field for databricks.sdk.service.supervisoragents.SupervisorAgent to no longer be required
Fixes (1) ›

  • Fixed Databricks CLI --profile fallback by detecting the CLI version at init time. The previous error-based detection was broken because --profile is a global Cobra flag silently accepted by old CLIs

Maintenance (2) ›
  • Detect Databricks CLI version at init time via databricks version, enabling version-gated flag support without additional subprocess calls

  • Validate Databricks CLI configuration at DatabricksCliTokenSource.__init__ time. Misconfiguration (missing profile and host, or --profile-unsupported CLI without a host fallback) now surfaces as IOError synchronously from construction rather than lazily from the first refresh() call. The exception type matches the previous refresh()-time behaviour, so callers who already catch IOError are unaffected

Read the original announcement →

https://github.com/databricks/databricks-sdk-py/releases/tag/v0.106.0

Related releases