Databricks Release Notes

New maintenance updates are available for supported Databricks Runtime versions. These updates include bug fixes, security patches, and performance improvements. For details, see:

• Databricks Runtime 18.0
• Databricks Runtime 17.3 LTS
• Databricks Runtime 17.2
• Databricks Runtime 17.1
• Databricks Runtime 16.4 LTS
• Databricks Runtime 15.4 LTS
• Databricks Runtime 14.3 LTS
• Databricks Runtime 13.3 LTS
• Databricks Runtime 12.2 LTS

Knowledge Assistant is now available in the following AWS regions: us-east-2, ca-central-1, eu-central-1, eu-west-1, ap-southeast-1 (requires cross-geo processing), and ap-southeast-2 (requires cross-geo processing).

Users in these regions can now use Knowledge Assistant to create a production-grade AI agent that can answer questions about their documents and provide high-quality responses with citations.

See Use Knowledge Assistant to create a high-quality chatbot over your documents.

Databricks Managed MCP servers is now Public Preview. Managed MCP servers allow your AI agents to securely connect to Databricks resources and external APIs. See MCPs.

You can now store MLflow traces in Unity Catalog tables using OpenTelemetry format and query them using Databricks SQL. This provides several benefits:

• Store unlimited traces in Delta tables for long-term retention and analysis
• Query trace data directly using SQL through a Databricks SQL warehouse
• Manage access control through Unity Catalog schema and table permissions
• Ensure compatibility with other OpenTelemetry clients and tools

See Store OpenTelemetry traces in Unity Catalog and Observe and find issues.

The standard Google Drive connector in Lakeflow Connect allows you ingest Google Drive files into Databricks. You can use read_files, spark.read, COPY INTO, and Auto Loader to create Spark DataFrames, materialized views, and streaming tables, enabling you to build custom pipelines for common file ingestion use cases. See Ingest files from Google Drive.

• Databricks Apps filtering: Genie One consumers can now filter to include or exclude inactive Databricks apps when browsing. Inactive apps are those with a status other than Running, such as Stopped, Deploying, or Crashed. See App status.
• Dashboard widget title search: Search now supports dashboard widget titles.
• Enhanced search rankings: Certified and favorite content is now boosted in search rankings in Genie One.
• Recents and favorites in discovery: Recents and favorites are now surfaced in the For you page.

• Unity Catalog lineage for Genie Agents (formerly Genie Spaces): Genie Agents (formerly Genie Spaces) are now discoverable in Unity Catalog lineage.
• Bulk actions for entity matching: Users can now make bulk actions for entity matching (add, remove, and refresh).

Lakebase is now generally available. GA includes autoscaling, scale-to-zero, instant branching, automated backups, point-in-time recovery, an adjustable storage quota, and expanded region availability.

See Get started with Lakebase Postgres.

You can now use AWS front-end PrivateLink for private connectivity to performance-intensive services like Zerobus Ingest and Lakebase Autoscaling. See Configure inbound PrivateLink for performance-intensive services.

Genie One is now generally available. Genie One is a simplified user interface designed for business users, providing a single, intuitive entry point to interact with data and AI in Databricks without requiring technical knowledge of compute resources, queries, models, or notebooks. Using Genie One, business users can view and interact with dashboards, ask data questions in natural language using Genie, and use custom-built Databricks Apps that combine analytics, AI, and workflows. See Use Genie One.

Serverless jobs now support custom base environments defined with YAML files for Python, Python wheel, and notebook tasks. For notebook tasks, you can either select a custom base environment in the job’s environment configuration or use the notebook’s own environment settings, which support both workspace environments and custom base environments.

For more information, see Manage workspace base environments.

Level of detail expressions let you control aggregation granularity independently of the dimensions in your visualizations. Use fixed level of detail expressions to aggregate over specific dimensions regardless of visualization groupings, or use coarser level of detail expressions to aggregate at a higher level by excluding dimensions. Common use cases include calculating percentages of total, comparing values to dataset-wide aggregates, and creating cohort-level metrics. See Level of detail (LOD) expressions.

The value sampling features in Genie have been renamed to better describe their function:

• Prompt matching (previously value sampling)
• Format assistance (previously example values)
• Entity matching (previously value dictionaries)

• Dashboard page navigation scroll bar: The horizontal scroll bar to navigate between dashboard pages is now thicker.
• Dashboard title character limit: Dashboard titles now have more horizontal space before truncating.
• Pivot table header height: Pivot table header height can be configured to enable wrapping for multi-row heights.
• Text widget markdown tables: Text widgets now support markdown table syntax.

• Mark assets as favorites: Click the star icon next to asset names in Genie One to mark them as favorites.
• Favorite and certified content filtering: Filter content by favorite and certified status in Genie One. See Listing pages.

• Databricks on AWS GovCloud support: Genie and Databricks Assistant are now generally available on Databricks on AWS GovCloud. See Databricks on AWS GovCloud.

Lakebase is now managed through the Lakebase App, accessed from the apps switcher in the Databricks UI. This replaces the previous workflow of navigating to the Compute tab in the Lakehouse UI.