Databricks releases, every cloud
- Tags for dashboards, notebooks, Databricks apps, and Genie Agents are now generally available
Tags support both governed and ungoverned key-value pairs. They can be applied to organize workspace assets.
- Service principal Genie usage is billed at standard pricing
It excludes service principals from the free monthly usage allowance.
- Scala UDFs can access Unity Catalog secrets
They use the caller's or function owner's permissions. Requires environment version 6 or Databricks Runtime 19.
- Scalar and Batch Unity Catalog Python UDFs are now generally available
They work on all supported compute types and remove the five-UDF-call-per-query limit. Environment version 6 or above is required for some capabilities.
- Time travel queries now enforce ABAC policies (Beta)
Requires Databricks Runtime 19 and Unity Catalog managed tables.
- Variant shredding will soon be enabled by default on new Variant tables Coming soon
Requires Databricks Runtime 17.3 or above. Breaks compatibility with Databricks Runtime 15.4.
- Run the change, do not just read it. Hands-on labs in your own Databricks workspace, graded when you submit. Browse labs
- Dashboard enhancements and bug fixes
The dataset results table and metric view preview grid now include features like column pinning and natural-language filtering.
- Genie Agents enhancements and bug fixes
PDF upload limit increased to 100 pages, and agents now only use attached sources.
- Metastore-level ABAC policies are in Beta
Applies across every catalog in the metastore. Replaces per-catalog policy setup.
- Skills in Genie One support multiple files and folders
Includes up to 3 folders deep and 50 files or folders per skill, limited to text files.
- Account-only user access to Genie One (Beta)
Eligible users are enrolled automatically, governed by Unity Catalog permissions.
- VARIANT cast extends to ARRAY and MAP columns in ABAC column masks
Requires Databricks Runtime 19 or above. Supports ARRAY and MAP in addition to STRUCT.
- Data Classification now scans Unity Catalog views (Beta)
Scans Unity Catalog views in addition to tables, using the same detection engine, and is currently in Beta.
- The READ METADATA privilege will soon be available by default for workspaces with the compliance security profile enabled Coming soon
Grants read-only visibility into object metadata, including security-sensitive information.
- The counter_diff SQL window function is now available
Calculates differences between consecutive cumulative counter values, available in Databricks Runtime 19 and above.
- Window measures on a numeric index column in metric views
Supports unitless offset and ranges on consecutive integer columns, useful for non-calendar periods like fiscal weeks.
- Dashboard enhancements
Supports more Tableau calculations and relationship-scoped filters.
- Databricks-provided MCP connectors for Genie One and Genie Code are generally available
Supports Google Drive, Gmail, and other tools, governed by Unity Catalog and Unity Gateway.
- Get this table by email. One Monday mail covering the week, filtered to the clouds and products you run. Weekly digest
- Genie Agents enhancements
The agent monitoring table now has a column picker and citations can point to specific pages in volume files.
- Genie One enhancements
Pinning a chat keeps it at the top of the sidebar. The mobile app supports MCP write actions.
- External secrets in Unity Catalog (Beta)
Connects Unity Catalog schema to AWS Secrets Manager, appearing as read-only objects.
- ABAC DENY policies are in Beta
Explicitly deny MANAGE ACCESS CONTROL privilege, taking precedence over grants.
- Databricks SQL version 2026.36 is rolling out in Current
It includes features from version 2026.32.
- See which chunks support each ai_search answer (Beta)
Added generate_citations option to show supporting chunks.
- time_bucket SQL function aligns timestamps to fixed-width time buckets
Replaces date_trunc for non-calendar units, requires origin and width parameters.
- Dashboard enhancements and bug fixes
It converts SQL datasets to local metric views and understands SQL dataset parameters.
- Genie Agents enhancements
Each Genie Agent now supports up to 50 data sources and stores up to 200,000 conversations.
- Genie One enhancements
Genie Code skills are on by default and can be customized. Backend steering allows sending messages while Genie One is answering.
- Web search in Genie One (Beta)
Searches public web for current info, citing sources as links.
- ABAC GRANT policies are generally available for models and services
They grant Unity Catalog privileges based on governed tags, replacing individual object grants.
- Databricks SQL version 2026.36 is now available in Preview
It includes features from version 2026.32.
- Unity Catalog Skills (Beta)
Publishes and governs agent skills, controlling access with Unity Catalog permissions.
- Dashboard enhancements and bug fixes
Embedded dashboards support filter APIs and rich text formatting for dynamic data references.
- Genie One enhancements
The Ask Genie button is hidden in embedded dashboards when unsupported. Push notifications are sent to iOS devices for scheduled tasks.
- Genie Agents enhancements
Genie Agents now display SQL function descriptions and use upgraded large language models.
- Genie Agents enhancements
Genie Agents now support chat mode and cancellable API responses. They also offer a downloadable visualizations API.
- Databricks SQL version 2026.32 is rolling out in Current
It is available in the Current channel and includes new features.
- Secrets in Unity Catalog will soon be available by default for workspaces with the compliance security profile enabled Coming soon
Replaces manual enablement, uses catalog.schema.secret namespace.
- The INSERT, UPDATE, and DELETE privileges are in Beta
Require Databricks Runtime 18.1 or above, alternative to MODIFY privilege.
- Context attributes in ABAC column mask and row filter policies (Beta)
Targets request context, such as OAuth client ID and on-behalf-of status, in Unity Catalog policies.
Headlines, dates and product areas are Databricks' own, and every item links to the note it came from. The one-line summaries are ours.