Databricks Release Notes

Databricks now snapshots your notebook’s Python variables before terminating idle serverless compute. When you reconnect, your notebook is automatically restored from its snapshot, letting you continue your work seamlessly.

See Automated session restoration for serverless notebooks.

To change the ownership of a share for Delta Sharing, you must now be the metastore admin. Share owners can no longer transfer ownership. See Update shares.

The Enhanced Security and Compliance add-on, which includes the compliance security profile, automatic cluster update, and enhanced security monitoring, is now available to all customers, with no contract changes required. You can now enable any of these features on a workspace to benefit from enhanced security capabilities. Now, only workspaces with enabled features incur charges, giving you flexibility and control over security spend. Customers with existing contracts that include the Enhanced Security and Compliance add-on are not affected by this change.

You can set the Databricks connector for Power BI to use the Arrow Database Connectivity (ADBC) driver instead of the ODBC driver.

See Configure ADBC or ODBC driver for Power BI.

Model Serving Foundation Model API pay-per-token endpoints and batch inference using Databricks-hosted models are now available in FIPS-compliant environments. This includes support for the following compliance standards offered through the compliance security profile:

• FedRAMP Moderate
• IRAP
• Canada Protected B

See Compliance standards with serverless and standard compute availability for region availability of these standards.

• Area chart error message fix: Fixed an error message that appeared when using disaggregated stacked area charts.
• Point map visualization UI update: Swapped the latitude and longitude sections in Point Map visualizations.

• Editable filters displays multiple values: Editable filters now display multiple values when Genie generates a SQL filter clause that includes more than one filter value.
• Edit space descriptions: You can now edit Genie Agent (formerly Genie Space) descriptions directly in place.
• SQL query recommendation: Genie now suggests popular SQL queries that come from notebooks. Previously, only popular SQL query objects were suggested.
• Faster entity matching identification: Entity matching (previously value dictionaries) now identifies unique values faster.

Databricks Apps now supports adding Lakebase database instances as app resources. You can then include these resources as environment variables and access them from your app code.

See Add a Lakebase resource to a Databricks app.

Tasks that wait on child processes (Run job and For each tasks) now have a separate limit for the number of tasks that can run simultaneously, and do not count against the overall limit.

See Resource limits.

• Dashboard import and export retains theme settings: When you import or export a dashboard as a file, theme settings are now preserved. To learn how to set a theme, see Theme settings.

• Dashboards now support up to 15 pages: The maximum number of pages per dashboard has increased from ten to fifteen. See Author dashboards.

• Annotation lines supported for temporal scales: You can now add annotation lines to charts with temporal scales.

• Parameterized schedules re-enabled: Support for parameterized schedules has been restored. See Create a schedule.

• Genie visualizations respect SQL sort order: Genie visualizations now default to the same sort order as returned by the SQL query.
• Permissions fix: Users with CAN MANAGE on a Genie Agent (formerly Genie Space) can now delete it.
• Datatype interpretation fix: Resolved an issue where string fields were incorrectly interpreted as date or time values.

Databricks SQL version 2025.20 is now available in the Preview channel. Review the following section to learn about new features and behavioral changes.

Use the UI to add and manage multiple Git credentials in the workspace, from one or multiple Git providers.

See Configure Git integration for Git folders.

Data lineage can now include external assets and workflows that are run outside of Databricks. If have workloads that run outside of Databricks (for example, first mile ETL or last mile BI), Unity Catalog lets you add external lineage metadata to augment the Databricks data lineage it captures automatically, giving you an end-to-end lineage view in Unity Catalog. This is useful when you want to capture where data came from (for example, Salesforce or MySQL) before it was ingested into Unity Catalog or where data is being consumed outside of Unity Catalog (for example, Tableau or PowerBI).

This external lineage metadata feature is in Public Preview. See External lineage.

The new Databricks JDBC Driver is now open source. The source code is publicly available under the Apache 2.0 license. Community contributions are welcome. See the Contribution Guidelines.

• Improved diff review for dashboard files: When committing dashboard changes to Git, queries and text box content now appear across multiple lines in the diff, making changes easier to review.

• Multi-statement queries in Genie on dashboards: Genie on dashboards now supports multistatement SQL queries.

• Improved schedule and subscribe workflow: You can now add subscribers when setting up a new schedule. Previously, you could only add subscribers to an existing schedule. See Create a schedule.

• New Genie API endpoints in Public Preview: You can now use the Genie API to list all conversations in a space, delete conversations, and delete Genie Agents (formerly Genie Spaces). See Genie in the REST API reference for details.

• More selective application of SQL functions in Genie: Genie's ability to apply SQL functions to relevant user questions has been improved.

MLflow now supports all compliance security profile standards supported by Databricks, such as HIPAA, PCI-DSS, and FedRAMP Moderate. For a full list of supported compliance standards, see Compliance security profile.

You can trigger job runs using file arrival triggers with file events on external locations, such as Amazon S3, Azure storage, or Google Cloud Storage. To benefit from this feature, the external locations must have file events enabled. File arrival triggers with file events allow a higher trigger limit (1,000 per workspace) and removes the 10,000 files per trigger limit. See File arrival triggers with file events.

Information Extraction and Custom LLM have a UI workflow to convert PDFs to Markdown and store the results in a table. Users can then use the resulting table as input in Information Extraction and Custom LLM.

See Use PDFs for Information Extraction and Custom LLM.