Metastore admins can use Unity Catalog permissions on the system.ai schema to decide which task-specific AI Functions people in the organization may run. The feature is generally available, but the page says to contact your Databricks account team to enable it.
The Databricks platform, as it moves
News
14 changes this week 6 shipped 8 in preview
150 items across the platformClear all
Top stories this week
Catalog
Release
Unity Catalog permissions on system.ai now control which AI Functions users can run (GA)
AI functions in SQLRead
AI Search
Release
AI Search can index Streaming Tables, Materialized Views and managed Iceberg tables directly
AI Search now accepts Streaming Tables, Materialized Views and managed tables using Iceberg v3+ as index sources, alongside Delta tables. You create a Delta Sync index pointing at the source table, so the intermediate Delta copy and custom synchronization pipeline are no longer needed. The post does not label the feature's maturity.
AI Search index types and sync modes +1Read
Jobs & Pipelines
Release
Lakeflow Jobs show job-level query performance metrics and insights (GA)
Improved Lakeflow Performance Observability is generally available. Lakeflow Jobs now show query performance metrics and insights at the job level for supported tasks, and an Optimize button opens Genie Code to help improve queries using those insights.
Monitoring runs: states, run history, trendsRead
Everything, day by day
Yesterday
- Unity Catalog permissions on system.ai now control which AI Functions users can run (GA)Metastore admins can use Unity Catalog permissions on the system.ai schema to decide which task-specific AI Functions people in the organization may run. The feature is generally available, but the page says to contact your Databricks account team to enable it.
- Default OAuth client IDs let ChatGPT, Claude, Cursor and others reach Databricks MCP serversChatGPT, Codex, Claude, Claude Code and Cursor can now connect to Databricks MCP servers, including the Genie One MCP server, with default OAuth client IDs. An account admin no longer has to create an OAuth app first. The feature is generally available.
- Unity Catalog volume files can be edited in the workspace file editor (GA)Files stored in Unity Catalog volumes can now be opened and edited in place in the workspace file editor, with no download and re-upload. You open the file from Catalog Explorer or from the schema browser in a notebook or the SQL editor. The feature is generally available.
Monday 26 October
- New account-level security features are available to all accounts as soon as they reach GAAccount-level security features, such as private connectivity, ingress and egress controls and encryption, no longer need an account admin to switch them on from the Previews page once they are generally available. Admins still have to configure each feature before it takes effect, and nothing changes in an account's security setup automatically.
- AI Search can index Streaming Tables, Materialized Views and managed Iceberg tables directlyAI Search now accepts Streaming Tables, Materialized Views and managed tables using Iceberg v3+ as index sources, alongside Delta tables. You create a Delta Sync index pointing at the source table, so the intermediate Delta copy and custom synchronization pipeline are no longer needed. The post does not label the feature's maturity.
- Lakeflow Connect adds a managed QuickBooks connector in BetaA managed QuickBooks connector is in Beta in Lakeflow Connect. It ingests accounting data from QuickBooks Online, such as accounts, customers, vendors, invoices, bills, payments and journal entries. It authorizes through OAuth user authorization and supports incremental ingestion with delete capture.
- Lakeflow Jobs show job-level query performance metrics and insights (GA)Improved Lakeflow Performance Observability is generally available. Lakeflow Jobs now show query performance metrics and insights at the job level for supported tasks, and an Optimize button opens Genie Code to help improve queries using those insights.
Friday 23 October
- Files API 3.0 for Unity Catalog volumes enters Beta with multipart uploadsFiles API 3.0 is in Beta for Unity Catalog volumes. It adds multipart uploads, so files larger than 5 GB can be uploaded, and conditional operations, which upload, download or delete a file only if its current version matches, to avoid overwriting unintended changes.
- Lakeflow Connect adds a Box Audit Logs connector in BetaA managed Box Audit Logs connector is now in Beta in Lakeflow Connect. It ingests enterprise audit events from Box into Databricks.
- DataFrame caching on serverless compute enters Public Previewdf.cache() and df.persist() now work on serverless compute, so an expensive result can be reused for the rest of a session. It needs serverless environment version 5 or above, plus Databricks Connect for Python 18.1.8 or Scala 18.1.5 or above. The preview is on by default for most workspaces and admins can switch it off from the Previews page.
Thursday 22 October
- Notebooks and files in one bundle can share serverless compute (Beta)Shared Serverless for Asset Bundles lets notebooks and files in the same bundle share one serverless compute and an environment defined in the bundle-root pyproject.toml. It is in Beta, and workspace admins enable it through Project-Scoped Shared Compute on the Previews page.
- Serverless job tasks can configure their environment from pyproject.toml (Beta)Serverless notebook and Python script tasks in Lakeflow Jobs can now take their job environment from a pyproject.toml file, set in the Jobs UI, the Jobs API or a bundle. The feature is in Beta, and workspace admins enable it through Project-Scoped Shared Compute on the Previews page.
- Lakeflow Connect adds a Microsoft Entra ID connector in BetaThe managed Microsoft Entra ID connector is now in Beta in Lakeflow Connect. It ingests users, groups, applications, service principals, directory roles, identity configuration and audit and sign-in logs. It authenticates with OAuth machine-to-machine and loads audit and sign-in logs incrementally.
Wednesday 21 October
- Genie Code tasks in Lakeflow Jobs are generally availableA Genie Code task runs a prompt autonomously as a step in a job, can read upstream task outputs and call tools, and returns a link to the resulting Genie Code chat. Starting in November 2026 it is available by default for workspaces with the compliance security profile enabled.
- Lakeflow system tables for pipelines reach general availabilityThe lakeflow system tables pipelines and pipeline_update_timeline are now generally available. They record the pipelines and pipeline updates in an account, and are documented in the jobs system table reference.
- Unity Gateway can screen MCP tool calls and results for jailbreaks (Beta)Unity Gateway now offers a Jailbreak guardrail for MCP traffic, in Beta. It screens MCP tool calls and their results for jailbreak and prompt injection attempts, and can be applied on input, output, or both.
- AI Runtime, serverless GPU training on Databricks, announced as generally availableDatabricks announced general availability of AI Runtime, its managed serverless GPU compute for training and fine-tuning models next to Unity Catalog data. It runs your own code, containers or Ray workloads from notebooks, IDEs or a CLI, logs runs to MLflow, and reports GPU usage in billing system tables. Workspaces with the compliance security profile are expected to get it later, around mid November.
Monday 19 October
- Asynchronous processing for Python transformWithState is generally availableThe Python row-based transformWithState API can now run state operations and user logic concurrently, which the release note says raises throughput with only minor code changes. The feature is generally available.
- Service principal OAuth tokens can carry a custom identity claim for per-user row filteringA custom identity claim can now be added to a service principal's OAuth token and used to filter rows in a dynamic view or an ABAC policy. This lets many end users be served through one service principal while each sees only their own rows.
- Databricks Runtime 19 hides row filter and column mask metrics in Spark ConnectIn the 19 October 2026 Databricks Runtime 19 update, Spark Connect DataFrame execution metrics no longer show internal operators and metrics tied to row filters and column masks, so policy execution details are not exposed through df.executionInfo.metrics. Query results do not change. Notebooks or tools that read detailed metrics inside protected subplans may need updating.
- DeepSeek V4.1 Flash will be retired on 19 November 2026
- DeepSeek V4.1 Flash (1017) is available on Unity Gateway with text and image inputA new DeepSeek V4.1 Flash version, labelled 1017, can now be called on Unity Gateway through Foundation Model APIs and accepts text and images. The release note points to a supported-regions list and says customers remain responsible for compliance with the applicable terms. It is also the recommended replacement for the retiring DeepSeek V4.1 Flash.
- Lakebase Insights is configured per project; observability-configuration schedules have stoppedLakebase Insights (Beta, formerly Lakebase Advanced Postgres Telemetry) is now set up in each project's Settings by users with CAN MANAGE, with on-demand runs, and no observability configuration is needed. Insights schedules created through an observability configuration have stopped and do not carry over; those configurations no longer write new telemetry to their Unity Catalog destinations. To keep scheduled insights, enable Insights and set an analysis frequency in each project.
Friday 16 October
- Saving an unchanged job no longer writes rows to the jobs system tablesSaving a Lakeflow job without changing its settings no longer adds rows to system.lakeflow.jobs or system.lakeflow.job_tasks, and change_time stays at the last real change. This applies to saves from the UI, the Jobs API and Declarative Automation Bundles.
- Genie file upload turns on by default for compliance security profile workspaces on 15 NovemberFrom November 15, 2026, file upload in Genie One and Genie Agents is available by default in workspaces that have the compliance security profile enabled, provided the workspace's compliance standard supports it. Workspaces on a standard without support still cannot use it; the feature availability table on each standard's page says which apply.
- File upload in Genie One and Genie Agents is GA; compliance-profile workspaces follow 15 NovemberFile upload in Genie One and Genie Agents is generally available, so workspace admins no longer need to turn on a preview. It is not supported in Genie One for users who have only consumer access. Starting November 15, 2026 it becomes available by default for workspaces with the compliance security profile, where the workspace's compliance standard supports file upload.
- Budgets for Unity Gateway and Genie accept ACLs with Budget Manager and Budget Viewer rolesBudgets now support access control lists. Account and workspace admins can give the Budget Manager or Budget Viewer role to users, groups or service principals, which delegates who can view and manage each budget. The change appears in both the Unity Gateway and the Genie release notes.
- Multi-statement transactions are GA on managed Delta and Iceberg tablesMulti-statement transactions are generally available for Unity Catalog managed Delta Lake and Apache Iceberg tables, so related changes across several tables commit or roll back together. The announcement pairs them with stored procedures, temporary tables and SQL scripting, and says they respect row filters and column masks.
- READ METADATA privilege now works with every compliance security profile standardThe READ METADATA privilege now supports all compliance security profile standards that Databricks supports. It is available by default for workspaces that have the compliance security profile enabled.
Thursday 15 October
- Data Classification detects 50 more classes and 16 more regionsData Classification now detects 50 new classes of sensitive data, covering payment card and financial data, personal and employee data, software secrets and regional identifiers. It also adds support for 16 more countries and regions. Regional tags are available in select regions only.
- MLflow run Logs tab for AI Runtime gains search, filters and downloadThe Logs tab on the MLflow run page for AI Runtime runs can now be searched and filtered by node, level and time range, follows a running workload, and lets you download logs. For new runs it replaces the log files that used to sit on the Artifacts tab, which are deprecated; logs are kept for 60 days, and logs of existing runs stay on the Artifacts tab.
- Lakebase Search tokenizer handles Chinese, Japanese and Korean textThe lakebase_tokenizer extension gains a morphological template that segments Chinese, Japanese and Korean text, with built-in tokenizer_chinese, tokenizer_japanese and tokenizer_korean configurations. It ships in lakebase_tokenizer 0.2.0; existing users run ALTER EXTENSION lakebase_tokenizer UPDATE and keep their configurations and indexes.
- Lakebase supports Postgres logical replication in and outNative Postgres logical replication now works with Lakebase in both directions. You can migrate an existing Postgres database into Lakebase and keep it in sync until cutover, or publish changes from Lakebase to external Postgres databases.
- Agentic code converter is GA for converting SQL dialects to Databricks SQLThe agentic code converter, powered by Genie Code, is generally available for SQL targets. It converts dialects such as T-SQL, Snowflake, Redshift, Oracle, BigQuery and Teradata inside a migration project, validates and fixes its output, and writes results to a workspace folder.
- Consumer access users can query Unity Gateway services without a preview toggleUsers holding the consumer access entitlement can now query shared model services and model provider services, and use Unity Gateway coding agent features, without an account admin enabling a preview first. The note warns this can raise model traffic, including to external providers; admins can set budgets and rate limits or disable direct model access.
- MySQL integrated CDC pipeline in Lakeflow Connect is generally availableThe MySQL integrated CDC pipeline has reached general availability. It handles extraction and applying changes in a single pipeline update, so unlike the gateway-based setup it needs no separate ingestion gateway.
- ADBC driver for the Power BI connector is GA and the default for new connectionsThe Arrow Database Connectivity (ADBC) driver for the Databricks connector for Power BI is now generally available. New connections in Power BI Desktop and the Power BI service use ADBC by default instead of ODBC; the docs describe how to pick either driver.
Wednesday 14 October
- Data Classification API and Terraform resource are generally availableThe Data Classification API and its Terraform resource have reached general availability. The auto-tagging configuration (auto_tag_configs) is still in Public Preview.
- Document Intelligence REST APIs for parse, extract and classify reach GASynchronous REST APIs for ai_parse_document, ai_extract and ai_classify are generally available, so an app or agent can process a file as it arrives. The ai_decide API, used for routing documents, is in Beta. Databricks Apps and agents built on Agent Bricks are the intended callers.
- Predictive AI Functions enter Beta: forecasting, anomalies and predictions in SQLPredictive AI Functions are in Beta: an upgraded ai_forecast plus new ai_top_drivers, ai_detect_anomalies, ai_predict_value and ai_predict_class, all callable from SQL. ai_forecast gains holiday support and external covariates. Genie can already call ai_top_drivers; the other four are listed as coming to Genie later.
- Private Network Gateway for serverless: Private Preview on AWS, Public Preview on AzurePrivate Network Gateway lets serverless workloads reach resources in your own network through one delegated subnet, configured in a Network Connectivity Configuration. On AWS it is in Private Preview, with enrollment through the account team; on Azure it is in Public Preview. Google Cloud support is described as in development.
Tuesday 13 October
- Clean rooms: one-step packaged clean room with an identity resolution provider is GAFrom the Clean Rooms page you can now create a ready-made clean room with an approved identity resolution provider, such as Acxiom, Adstra, LiveRamp or The Trade Desk, in a single step. The provider is already set as the package provider, so you add your data and run its analysis straight away.
- Clean rooms: shared output schema is GAClean room runs can write to a shared output schema that every collaborator can read as each table is created, even before the run finishes. It is meant for sharing results, logs and debugging information.
- Clean rooms: JAR workloads are GAClean rooms can now run compiled Java or Scala code. A JAR runs as its own clean room analysis, or a function registered from a JAR can be called as a temporary UDF from a clean room notebook.
- External Spark clients can write to Unity Catalog managed Delta tables (GA)Creating and writing Unity Catalog managed Delta tables from external Apache Spark clients is now generally available. The tables must have catalog commits enabled.
- External engines can query Unity Catalog views via the Unity REST API (GA)External engines can now read Unity Catalog views defined over Delta Lake tables through the Unity REST API, and the capability is generally available. Teams that expose governed data to non-Databricks engines can include views in that access path instead of only tables.
- On-demand state repartitioning is generally available for stateful streamingYou can now change the number of shuffle partitions of a stateful Structured Streaming query or a Lakeflow pipeline without losing its checkpoint state. The feature is generally available.
- Gemini 4 Argon is now available on Unity Gateway through Foundation Model APIsGoogle's Gemini 4 Argon model can now be reached through Unity Gateway using Foundation Model APIs. The release note says customers remain responsible for complying with the applicable model terms.
Monday 12 October
- Private Link from serverless compute to AWS-managed resources is generally availablePrivate connectivity from Databricks serverless compute to AWS-managed resources such as S3 buckets, previously in Public Preview, is now generally available. Traffic stays inside the AWS network instead of crossing the public internet.
- Agent memory pipelines distil conversations into agent memory (Beta)A memory pipeline links a session store to a memory store. Each run uses an LLM to extract durable facts, preferences and decisions from stored conversations into memory entries that an agent can recall later. Pipelines run on demand or about every 24 hours, and the feature is in Beta.
- Governed tags can now propagate through lineage to derived tables (Beta)With tag propagation enabled on a governed tag, Unity Catalog uses lineage to apply that tag to derived tables and views after supported operations such as CREATE TABLE AS SELECT, CREATE VIEW, INSERT INTO ... SELECT and MERGE. The aim is consistent classification and ABAC enforcement downstream, at table and column level. It is in Beta.
- Unity Gateway Skills REST API enters Beta with CLI and SDK supportA REST API for managing skills on Unity Gateway is now in Beta, covering create, read, update, list, delete and finalize. The Databricks CLI and the Python, Go, Go Modular, Java and JavaScript SDKs support it.
- Lakeflow pipeline parameters are generally available for SQL sourcesKey-value parameters for Lakeflow pipelines have left preview. One pipeline's SQL source can run against different catalogs, schemas or tables: defaults live in the pipeline settings, an update can override them, and the SQL refers to them with named parameter syntax.
- Feature Views reach general availability for batch, streaming and request-time dataFeature Views, the declarative way to define and compute features, are generally available for batch, streaming and request-time data. For workspaces with the compliance security profile, they will be switched on automatically in mid-November 2026.
Friday 9 October
- Databricks Apps: workspace admins can set default Python and npm package repositoriesWorkspace-level package repositories for Databricks Apps are generally available. Admins set default Python and npm repositories that every app inherits, and an app can still override them. A separate upcoming note says that from November 13, 2026 the feature is turned on automatically for workspaces with the compliance security profile.
- JDBC Unity Catalog connections to be on by default under the compliance profileIn mid-November 2026, JDBC Unity Catalog connections, including SSL certificate support, will be available by default for workspaces with the compliance security profile enabled. Such a connection reaches an external database through a custom JDBC driver, with access governed by Unity Catalog. The note gives a month, not a day.
- Unity Gateway model provider services can reach Amazon Bedrock Mantle modelsAmazon Bedrock Mantle models can now be called through Unity Gateway model provider services. The calls are governed by Unity Catalog permissions, rate limits and budgets, and are recorded in inference tables.
- Genie One can run code in an isolated sandbox (Public Preview)Genie One can now run code in a secure, isolated sandbox for analyses that go beyond SQL. The code runs with the user's existing permissions, so Unity Catalog governance and data access controls still apply. The feature is in Public Preview.
- OpenJev can be the evaluator model for LLM-as-a-judge service policies (Beta)In Beta, a custom LLM-as-a-judge service policy on Unity Gateway can use OpenJev (Qwen3.5 4B) as its evaluator model service. The page says OpenJev classifies content without generating output tokens, so it typically responds faster than a chat evaluator.
- Unity Gateway service policies can require caller-supplied request tags (Beta)In Beta, a custom service policy on Unity Gateway can check tags that the caller attaches to a request, for example requiring a given tag before a model request or an MCP tool call goes through. This lets a policy decide on caller context rather than only on content.
Thursday 8 October
- Google Workspace connector for Lakeflow Connect reaches BetaLakeflow Connect has a managed Google Workspace connector in Beta. It ingests audit activity from Google Workspace applications and services into Databricks, using OAuth user authorization and incremental ingestion.
- Productboard connector for Lakeflow Connect is in BetaLakeflow Connect has a managed Productboard connector in Beta. It ingests product hierarchy, roadmap, objectives, feedback notes and workspace members, using Server-to-Server OAuth and incremental ingestion for notes. Workspace admins control access from the Previews page.
- Unity Catalog Scala, Java and Python UDFs to be on by default under the compliance profileUnity Catalog Scala and Java UDFs, and scalar and batch Unity Catalog Python UDFs, will be available by default in November 2026 for workspaces with the compliance security profile enabled. The note gives the month only, not a day.
- Genie One file upload accepts plaintext files such as .txt and Markdown (Beta)File upload in Genie One now also takes plaintext files, including .txt and Markdown (.md), alongside CSV, Excel, PDF, image, Word and PowerPoint files. The release note labels this addition Beta; it does not announce general availability.
- The Discover page and domains in Unity Catalog reach general availabilityDiscover, a curated way to browse assets without knowing the catalog hierarchy, and domains with subdomains are now generally available. Domains group tables, dashboards, notebooks and Genie Agents by business area, allow curated highlights, and surface AI-recommended assets.
- Workday Data Connect federation connector for Unity Catalog opens in BetaA Lakehouse Federation connector lets Unity Catalog read Workday's shared Iceberg tables in place, with no ingestion pipeline and read-only access. It needs Databricks Runtime 19 or above, and a workspace admin must enable it from the Previews page. Lakeflow Connect for Workday remains the route when a durable copy is needed.
Wednesday 7 October
- DeepSeek V4 Flash (0731) will be retired on 5 November 2026
- Context-based ingress policies for inbound Private Link are generally availableContext-based ingress policies for inbound Private Link have reached general availability. They control which registered VPC endpoints can reach workspaces and combine endpoint rules with identity and request-type conditions. If the feature is not yet enabled for an account, an account admin can turn it on from the account console Previews page.
- AI/BI dashboards can pre-compute dataset results on a schedule (Beta)Dataset materialization for AI/BI dashboards is in Beta. Dataset results are computed on a schedule and stored, so a published dashboard renders from the stored results instead of rerunning expensive queries on every load or interaction.
- Unity Gateway external service policies send traffic to a partner guardrail (Beta)In Beta, an external service policy on Unity Gateway forwards request and response content to a third-party guardrail vendor and blocks the call if the vendor denies it. Applications need no changes.
- On-behalf-of-user authorization for Databricks Apps is generally availableDatabricks Apps can now call supported APIs as the signed-in user, so Unity Catalog row filters and column masks apply to that user, while API scopes cap what the app may do. A read-only query app can request sql:restricted-query rather than the broader sql scope, and workspace admins can restrict which scopes developers may add.
- DeepSeek V4.1 Flash gains priority pay-per-token and reserved throughputDeepSeek V4.1 Flash can now be served with priority pay-per-token for latency-sensitive traffic, or with reserved provisioned throughput for dedicated capacity.
- Claude Haiku 5.5 is available on Unity Gateway through Foundation Model APIsAnthropic's Claude Haiku 5.5 can now be used on Unity Gateway via Foundation Model APIs. The note reminds customers that they are responsible for complying with the applicable terms.
Tuesday 6 October
- CREATE OR REPLACE VIEW will keep grants, tags and comments from 2 December 2026From December 2, 2026, CREATE OR REPLACE VIEW will update a view in place instead of dropping and recreating it, so grants, tags and comments are preserved. Governed tags on a view column must be removed before the column is dropped this way, otherwise the statement fails with CANNOT_DROP_TAGGED_COLUMN. Adding NOCOPY drops grants, tags and comments explicitly, though governed tags are still retained.
- Genie Code CLI, a terminal coding agent for Databricks, is in BetaGenie Code CLI is a coding agent that runs in a terminal and works on local files. It uses the Databricks CLI to find data, answer questions, and build and deploy pipelines, models and apps. Model access goes through Unity Gateway. It is labelled Beta.
- Window measures in metric views reach general availabilityMetric views can now define window measures, which cover rolling windows, cumulative totals and semiadditive aggregations such as moving averages, period-over-period change and running totals. The feature is generally available as of this note. Anyone modelling time-based KPIs in metric views no longer has to treat it as a preview.
Monday 5 October
- Databricks Runtime 10.4 LTS ends on 2026-11-01
- Genie One MCP Beta endpoint (/api/2.0/mcp/genie) ends on 2026-10-31
- DeepSeek V4 Pro (0813) ends on 2026-10-30
- ai_parse_document now returns chart values as structured JSONThe ai_parse_document function extracts chart values as structured JSON and describes each chart's series, legends and axes. Search and agents can then use those values when answering.
- Genie One skills can be shared and governed through Unity Gateway (Beta)A skill built in Genie One can now be published and shared with a team in Beta. Access is governed by Unity Gateway and uses ordinary Unity Catalog privileges on the catalog and schema that hold the skill.
- Lakeflow Jobs can use up to five triggers per job (Beta)A Lakeflow job can now carry up to five triggers. Trigger types can be mixed, several triggers of the same type are allowed, and each trigger can be paused on its own. The release note labels this Beta.
- NEAREST BY: a SQL join for batch vector search in Databricks RuntimeNEAREST BY is a SQL join that finds, for each row on the left, the k closest rows on the right by vector similarity or distance, either EXACT or APPROX. It runs in Photon with a fused operator, and APPROX queries can use an IVF index stored as a liquid-clustered Delta table. The post does not state a release stage, so availability needs checking against the documentation.
- Iceberg REST catalogs gain UNREGISTER; Unity Catalog support is in private previewDatabricks contributed an UNREGISTER endpoint to the Apache Iceberg REST catalog specification. It removes a table from one catalog without touching its files and returns the latest metadata location, which another catalog then takes over with REGISTER, so a table has one managing catalog throughout the move. REGISTER and UNREGISTER on Unity Catalog are in private preview and available through the account team.
Friday 2 October
- Lakeflow pipelines can run as an account groupA Lakeflow pipeline's run-as identity can now be an account group, in addition to a user or a service principal. Pipeline updates run with the group's permissions, so a team can own an automated pipeline without it depending on one person's account.
- Customer-managed keys can now decrypt query history text fields (GA)A customer-managed key set on the system catalog can now be used to decrypt the statement_text and error_message fields in the query history system table. The release notes mark this generally available as of 2 October 2026. Teams that encrypt query history with their own key should read the docs section on reading encrypted fields.
- Gemini 2.5 Pro and 2.5 Flash retired; Gemini 3.1 Pro or 3.5 Flash replace them
- OpenJev (Qwen3.5 4B) is available on Unity GatewayOpenJev (Qwen3.5 4B) is now offered on Unity Gateway as a pay-per-token model service. The release notes say it is meant for yes/no, choice and scoring questions that return structured answers with probabilities, and that customers must check compliance with the applicable model terms.
- Standard ai_extract moves to a faster model checkpointThe standard (default) mode of ai_extract now runs on a new model checkpoint that the release notes describe as 2.3x faster, with better overall quality and no extra cost. Queries that already use the default mode pick up the change without edits.
- JDBC Unity Catalog connections are generally availableJDBC connections in Unity Catalog have reached general availability on serverless compute, Databricks Runtime 19 and above, and SQL warehouses 2025.40 and above, with SSL certificate support added. OAuth M2M authentication for these connections is still in Beta. Anyone who read the earlier Public Preview material should treat the GA scope above as the current one.
Thursday 1 October
- Lakeflow Connect adds a managed GitLab connector in BetaLakeflow Connect now has a managed GitLab connector in Beta, ingesting repository, project, group, CI/CD and security data. It uses OAuth U2M authentication and supports incremental ingestion, UI and API pipeline authoring, and Declarative Automation Bundles. Workspace admins control access from the Previews page.
- Cross-engine ABAC is generally available for external enginesSupported external engines can now read Unity Catalog managed tables while ABAC policies, row filters and column masks are enforced by Databricks serverless compute. The feature is generally available; requirements, supported clients and serverless compute costs are listed on the linked documentation page.
- LTAP Direct Writes for Lakebase synced tables is generally availableLTAP Direct Writes, which writes the initial load and full refreshes of a Lakebase synced table straight to the storage layer instead of through the compute endpoint, has left Beta. You opt in per synced table under Synchronization settings, and it works on Lakebase projects running Postgres 16, 17 or 18.
- Built-in IP and CIDR functions are generally available on Databricks Runtime 18 LTS+Databricks now ships native SQL, PySpark and Scala functions for IPv4 and IPv6 addresses and CIDR blocks: parsing, canonicalizing, inspecting, converting to binary, and ip_cidr_contains for containment joins. try_ variants return NULL on bad input instead of failing. They are generally available on Databricks Runtime 18 LTS and above.
Wednesday 30 September
- Excel and Google Sheets writeback now needs a workspace admin settingWriting data back to Databricks from the Excel add-in or from Google Sheets now requires a workspace admin to turn on the "Allow spreadsheet writeback" security setting. Workspaces where it stays off will see writeback stop working, so admins who have spreadsheet users should check the setting.
- ai_decide enters Beta: fast structured decisions over text and data in SQLai_decide is a new AI Function in Beta that answers one or more questions about an input in a single call. For each question it returns a probability, one label from criteria you define, or a score on an ordered scale. The input can be plain text, a JSON object or array, or the VARIANT output of ai_parse_document or ai_extract. It runs from SQL for batch work and through a REST API for real-time use; the SQL function needs Databricks Runtime 15.4 LTS or above and is not available on Databricks SQL Classic. Typical uses are tagging, routing prompts to a model and judging agent answers.
- Query tags for SQL warehouses reach general availabilityQuery tags are now generally available. You attach custom key-value tags to SQL workloads to group, filter and attribute cost on Databricks SQL warehouses and Lakehouse//RT compute. The tags show up in system.query.history and on the Query History page.
- Table update triggers in Lakeflow Jobs now cover OpenSharing objects and system tables (GA)Lakeflow Jobs table update triggers can now watch OpenSharing objects and system tables, and this is generally available. A job can start when a shared object or a system table changes, instead of on a schedule. Anyone polling these tables on a timer can consider switching to a trigger.
Tuesday 29 September
- Genie One memory is generally availableGenie One can keep facts you give it, such as preferences and reporting conventions, and apply them in later conversations. Memories are private to each user and are not shared; you can ask Genie One to remember something or accept a suggestion it makes.
- Agent Bricks CLI enters Beta for building and deploying custom code agentsThe Agent Bricks CLI (databricks-agentbricks) is a command-line tool in Beta for custom code agents. It scaffolds a project from a bundled framework template, runs it locally and deploys it to the Databricks agent runtime, with managed memory, sessions, tools and MLflow tracing set up from one authenticated command.
- OpenAI GPT-6.1 Sol added to Unity Gateway, reachable through Foundation Model APIsGPT-6.1 Sol from OpenAI can now be called through Unity Gateway using Foundation Model APIs. The release note adds that customers remain responsible for complying with the applicable terms of the model provider.
Monday 28 September
- Lakebase Search gets lakebase_tokenizer for configurable keyword matchingA new lakebase_tokenizer extension controls how Lakebase Search splits text into terms: normalization and accent stripping, your own stopword and synonym lists, and stemming. It produces standard tsvector values, so it works with to_tsvector, GIN indexes and the lakebase_bm25 index; install it after enabling Lakebase Search in project settings.
- xAI Grok 4.7 added to Unity Gateway as a pay-per-token Foundation ModelGrok 4.7 from xAI is now available on Unity Gateway and can be called through pay-per-token Foundation Model APIs. Customers remain responsible for complying with the applicable terms.
- Lakebase Search is GA: BM25 and vector search extensions in PostgresLakebase Postgres now ships two search extensions, lakebase_vector for approximate nearest-neighbour search and lakebase_text for BM25 full-text search, generally available on AWS and Azure. Together they allow semantic, keyword and hybrid search in one SQL query next to operational tables, without a separate search system. The vendor's own benchmarks compare it favourably with pgvector; the post also says Databricks AI Search remains the better fit for managed retrieval with little tuning.
- Lakeflow Connect adds a 1Password Event Logs connector in BetaThe 1Password Event Logs connector is in Beta in Lakeflow Connect. It uses the 1Password Events API to ingest audit events, item usages and sign-in attempts, and authenticates with an Events Reporting bearer token. Workspace admins control access from the Previews page.
- Lakeflow Connect adds an Akamai WAF connector in BetaThe Akamai WAF connector is in Beta in Lakeflow Connect and ingests web application firewall security events from Akamai. It authenticates with Akamai EdgeGrid API client credentials. Workspace admins control access from the Previews page.
- Lakeflow Connect adds a CrowdStrike Falcon Event Stream connector in BetaThe CrowdStrike Falcon Event Stream connector is in Beta in Lakeflow Connect and ingests Falcon Event Stream events. It authenticates with OAuth 2.0 client credentials from a CrowdStrike API client that has Event streams Read permission. Workspace admins control access from the Previews page.
- Lakeflow Connect adds a Workday Activity Logging connector in BetaThe Workday Activity Logging connector is in Beta in Lakeflow Connect. It incrementally ingests Workday user activity log events and authenticates with an OAuth refresh token. Workspace admins control access from the Previews page.
- Lakeflow Connect adds a Microsoft 365 Unified Audit Logs connector in BetaA managed Microsoft 365 Unified Audit Logs connector is in Beta in Lakeflow Connect. It ingests unified audit events from Microsoft Entra ID, Exchange, SharePoint and other Microsoft 365 workloads, uses OAuth machine-to-machine authentication, and loads incrementally.
- Lakeflow Connect adds an Okta System Logs connector in BetaA managed Okta System Logs connector is in Beta in Lakeflow Connect. It incrementally ingests audit and security events from the Okta System Log, authenticates with an SSWS API token, and supports Declarative Automation Bundles and API-based pipeline authoring.
- Thinking Machine Labs Inkling will retire on October 30, 2026
Saturday 26 September
- Horizontal scaling for Databricks Apps is now GAA Databricks app can now run across multiple instances behind one app URL instead of just one, for higher availability and concurrency. Deployments get zero-downtime updates and best-effort session affinity between instances. This horizontal scaling capability has reached general availability.
- How to set up a governed JDBC connection in Unity CatalogWalks through Databricks' new Unity Catalog JDBC connection type (Public Preview on Databricks Runtime 18.1 and DBSQL 2025.40 and above): staging the driver JAR in a Unity Catalog volume, creating the connection as a securable object, and granting access with GRANT USE CONNECTION so callers never see the URL or credentials. Compares it against query federation and PySpark data sources so a reader can pick the right one for a given external source.
Friday 25 September
- Deleting a Unity Catalog pipeline now keeps its tables unless you cascadeDeleting a Unity Catalog pipeline in the default publishing mode, from the UI or the Pipelines DELETE API, now leaves its materialized views, streaming tables and views in place. They can still be queried but cannot be updated until they move into an active pipeline. To remove the tables too, the API request must set the cascade field to true; Hive metastore pipelines are unaffected.
- Genie Code can now build document processing pipelinesGenie Code can now orchestrate an end-to-end document processing pipeline, handling parsing, classification, extraction, and validation in one flow. Pointed at unstructured PDFs, contracts, or invoices, it checks extraction quality against labeled data and iterates to improve it.
Thursday 24 September
- Tags for dashboards, notebooks, apps, and Genie Agents are GATags can now be applied to dashboards, notebooks, Databricks apps, and Genie Agents to organize workspace assets. The release is generally available and supports both governed and ungoverned key-value tags.
- ai_search can query Lakebase synced tables, in BetaThe ai_search function can now retrieve from Lakebase synced tables that have Lakebase Search enabled, as well as from AI Search indexes. A call adds a lakebase_search knowledge source, and it can be combined with vector_search sources in the same call. The release notes label this Beta.
Wednesday 23 September
- Lakebase writes its Postgres telemetry to system.lakebase tables (Beta)In Beta, Lakebase records query plans, session history, wait events, schema-change history, health data and Postgres logs into the system.lakebase system tables with no setup. Records are kept for 7 days, and account and metastore admins can read the tables and grant access to others.
- DeepSeek V4 Pro (0813) will retire and be replaced by V4.1 Flash
- Time travel can enforce ABAC policies on qualifying managed tablesIn Databricks Runtime 19, time travel can be used on qualifying Unity Catalog managed tables while ABAC row filter and column mask policies are enforced. The feature is in Beta, so learners should treat it as preview behaviour rather than generally available behaviour.
- Governed agentic app building is in betaDatabricks says you can now build governed data and AI apps without code. Genie App Builder authors apps in natural language, App Spaces enforce governance, and Serverless Micro Apps scale to zero when idle.
- Scala UDFs can now read Unity Catalog secretsSession-scoped and Unity Catalog Scala UDFs can now access secrets managed in Unity Catalog. Session-scoped UDFs use the caller's permissions, while Unity Catalog Scala UDFs declare secrets in a SECRETS clause and use the function owner's permissions. Users must check the stated environment, Databricks Runtime, access mode, and SQL warehouse requirements before using this in a UDF.
- Unity Catalog Python UDFs are GA for scalar and batch useScalar and batch Unity Catalog Python UDFs are now generally available on all supported compute types. The release adds named handlers for row-at-a-time scalar UDFs, removes the five-UDF-call-per-query limit, and aligns scalar TIMESTAMP inputs with PySpark behavior. Some capabilities require environment_version 6 or above, and requirements vary by feature and compute type.
Tuesday 22 September
- Variant shredding will be added by default to new VARIANT tablesStarting November 18, 2026, Databricks will automatically add the variantShredding table feature to new tables with a VARIANT column when they are created on Databricks Runtime 17.3 or above. Variant shredding stores frequently accessed fields in a typed, columnar layout to improve query performance. Clusters on Databricks Runtime 15.4 cannot read from or write to those tables, so affected clusters should be upgraded to Databricks Runtime 16.4 or above before the change takes effect.
- Coding agents can be configured centrallyOrganizations can now centrally configure coding agents, including models, MCP servers, skills, a default harness and models, and Smart Routing. The settings apply when an agent starts, or can be rolled out through mobile device management.
- Unified tracing includes coding-agent activityUnity Gateway adds coding-agent activity to unified tracing through the Unity Gateway CLI. Traces can show the tools and skills used alongside model requests in Govern > Traces or with Genie.
- ug CLI connects coding agents to Unity GatewayThe Unity Gateway CLI (ug) can connect Claude Code, Codex, Pi, and other coding agents to Unity Gateway with one command. It handles authentication and configuration, adds approved MCP servers and skills, and shows spend in the terminal.
- Unity Catalog Pages added as a home for Genie Ontology knowledgeDatabricks describes Unity Catalog Pages as a governed home for business knowledge used with Genie Ontology. Readers working with Genie Ontology or Unity Catalog Pages should treat Pages as part of where business context for agents is managed.
- OpenAI GPT-6 Luna is available on Unity GatewayOpenAI GPT-6 Luna is now available on Unity Gateway. It can be accessed through Foundation Model APIs pay-per-token, and customers are responsible for complying with applicable terms.
- OpenAI GPT-6 Sol is available on Unity GatewayOpenAI GPT-6 Sol is now available on Unity Gateway. It can be accessed through Foundation Model APIs pay-per-token, and customers are responsible for meeting the applicable terms.
- Lakebase backup schedules can now be managed through a Beta APIDatabricks added a Beta snapshot schedule API for Lakebase. It lets you read and set a root branch's automated snapshot cadences, including daily, weekly, or monthly schedules, and the retention period for each cadence.
- Shopify connector is in Beta for Lakeflow ConnectLakeflow Connect now has a Shopify connector in Beta. It ingests Shopify e-commerce data, including products, orders, customers, inventory, fulfillments, and discounts, into Databricks, and workspace admins can manage access from the Previews page.
Monday 21 September
- Genie One MCP beta endpoint is deprecated for a Unity Gateway service
- Model provider service API keys can use Unity Catalog secretsA model provider service API key can now be backed by a customer-owned Unity Catalog secret. The service reads the secret at query time, so rotating the secret takes effect without changing the service.
- system.query.history is generally availableThe system.query.history system table is now generally available. It provides account-wide records for queries run on SQL warehouses or serverless compute, including execution status, duration metrics, and query source.
Friday 18 September
- Lakebase Search is generally availableLakebase Search is now generally available for Lakebase projects. It uses the lakebase_vector and lakebase_text Postgres extensions for vector search and BM25 full-text search in the same database, instead of using a separate search stack. Lakebase users who want this must enable Lakebase Search in project settings and install the extensions.
- Databricks Apps telemetry is generally availableConfiguring telemetry for Databricks Apps is now generally available. It collects traces, logs, and metrics using the OpenTelemetry protocol and stores them in Unity Catalog tables. After telemetry is enabled, Databricks captures system logs and usage events automatically, and app developers can add custom instrumentation with an OpenTelemetry SDK.
Thursday 17 September
- Base environments are in beta on classic computeBase environments can now be used to manage Python dependencies on classic compute. The beta supports both Databricks-provided environments and custom workspace-level environments.
- Moonshot AI Kimi K2.7 will retire and Kimi K3 replaces it
- ABAC policies can now apply at the metastore level in BetaUnity Catalog now has Beta support for attaching an ABAC policy at the metastore level. A row filter, column mask, GRANT, or DENY policy attached there applies across every catalog in the metastore, including catalogs created later, instead of being replicated per catalog.
- Genie One skills now support multiple files and foldersGenie One skills can now include multiple files and sub-folders. Each skill can use folders up to three levels deep and up to 50 files or folders, and skills can load text files only.
Wednesday 16 September
- A worked example of what Spark Declarative Pipelines solves, with F1 dataA practitioner walkthrough that builds a small pipeline over Formula 1 race data to show what Spark Declarative Pipelines (SDP) automates: resolving dependency order from declared datasets instead of hand-built DAGs, retries, state management and an event log built into the framework, and triggered runs that reuse Structured Streaming's micro-batch engine so the same declarative pipeline handles batch and streaming without separate code paths.
- Managed agent memory and sessions are now in BetaDatabricks added Beta managed agent memory and managed agent sessions, two fully managed Lakebase-backed stores for agent state that work with agents built on any framework. Memory persists across conversations and supports semantic search; sessions store conversation history for the agent to read at the start of a turn and append to while it runs.
- ABAC column masks can use VARIANT functions for ARRAY and MAP columnsABAC column masks can now use a single VARIANT function for ARRAY and MAP columns, in addition to STRUCT and other column types. This requires Databricks Runtime 19 or above, so readers using masks on ARRAY or MAP columns need to check their runtime before using it.
- Anthropic models removed from AWS GovCloud IL5 workspaces
- Unused OAuth client secrets are deleted after 90 daysDatabricks now deletes a service principal's OAuth client secret after 90 days without use. If that deleted secret is still needed, create a new OAuth secret.
- Unity Gateway API and developer tools are generally availableThe Unity Gateway API is generally available for managing model services, model provider services, and MCP services. Create, read, update, list, and delete operations are supported across Terraform, the Databricks CLI, and the listed Databricks SDKs; Declarative Automation Bundles support through the CLI is in Beta.
Nothing of that kind yet.