Databricks Release Notes

The Akamai WAF connector is now available in Lakeflow Connect in Beta. Workspace admins can control access to this feature from the Previews page.

The connector ingests web application firewall (WAF) security events from Akamai into Databricks. Authentication uses Akamai EdgeGrid API client credentials.

See Akamai WAF connector.

The 1Password Event Logs connector is now available in Lakeflow Connect in Beta. Workspace admins can control access to this feature from the Previews page.

The connector uses the 1Password Events API to ingest audit events, item usages, and sign-in attempts into Databricks. Authentication uses an Events Reporting bearer token.

See 1Password Event Logs connector.

Databricks SQL version 2026.39 is now available in the Preview channel.

Databricks SQL version 2026.38 is rolling out to the Current channel. See features in 2026.38.

The system.ai_gateway.usage system table now includes coding-agent session metadata and invocation source. New token-detail fields capture tokens written to the 5-minute and 1-hour prompt caches, along with file and web search counts. See Track model usage.

You can now apply tags to dashboards, notebooks, Databricks apps, and Genie Agents to organize and categorize your workspace assets. This capability is now generally available. Both governed and ungoverned key-value tags are supported. See Apply tags to Unity Catalog securable objects.

When you delete a Unity Catalog pipeline that uses the default publishing mode, the UI and the Pipelines DELETE API retain associated materialized views, streaming tables, and views by default. The retained tables can be queried but cannot be updated until moved into an active pipeline. To delete a pipeline and its tables together, set the cascade field to true in the API request. This change does not affect Hive metastore pipelines.

See Delete a pipeline and Delete a pipeline.

The Genie One MCP server is now generally available as a Databricks-provided MCP Service, system.ai.genie_one_mcp, in Unity Gateway. It exposes Genie as a conversational tool over the Model Context Protocol (MCP), so any MCP client or agent, such as Claude, ChatGPT, or Cursor, can ground its data questions in Genie's trusted insights, with Unity Catalog permissions enforced on every request. See Genie One MCP server.

The previous Beta endpoint, https://<workspace-hostname>/api/2.0/mcp/genie, is deprecated and will be sunset on October 31, 2026. Move all workloads to the system.ai.genie_one_mcp MCP Service before then.

Genie Code can now build end-to-end document processing pipelines by orchestrating parsing, classification, extraction, and validation. Point it at unstructured PDFs, contracts, or invoices, and it evaluates extraction quality against labeled data and iterates to improve it. See Use Genie Code for intelligent document processing.

• Monitoring table column picker: You can now choose which columns appear in the agent monitoring table on the Monitor tab. See Monitor the agent.
• Page-level citations for volume files (Beta): When a Genie Agent answers using content retrieved from an attached volume, its citations can now point to the specific page of the source file. See How a Genie Agent analyzes files.

Genie usage by service principals is now billed at standard Genie pay-as-you-go pricing. Service principals are excluded from the free monthly usage allowance that applies to individual users. See Manage budgets and cost controls for Genie and Monitor and understand your Genie cost.

Sharing metric views with OpenSharing is now generally available. Add metric views to a share and query them with MEASURE(...) on serverless compute. See Add metric views to a share and Read shared metric views.

The ai_search function can now retrieve from Lakebase synced tables with Lakebase Search enabled, in addition to AI Search indexes. Add a lakebase_search knowledge source to search a Lakebase synced table, or combine it with vector_search sources in the same call. This feature is in Beta.

See ai_search function.

Scalar and Batch Unity Catalog Python UDFs are now generally available on all supported compute types. This 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. You can also install custom dependencies and access Unity Catalog service credentials and secrets from your functions. Some capabilities require the UDF definition to explicitly set environment_version to 6 or above, including the newly generally available scalar capabilities on serverless compute. Requirements vary by feature and compute type. See Python UDF feature requirements.

You can now build governed data and AI apps without code. Genie App Builder builds apps from natural-language prompts inside an App Space that enforces governance, and the apps scale to zero when idle. See Genie App Builder.

DeepSeek V4 Pro (0813) will be retired on October 30, 2026. DeepSeek V4.1 Flash is the recommended replacement. See Deprecated and retired models for migration guidance.

This feature is in Beta. Starting in Databricks Runtime 19, you can use time travel on qualifying Unity Catalog managed tables while enforcing ABAC row filter and column mask policies. See Time travel with ABAC policies.

Lakebase captures its Postgres observability telemetry into the system.lakebase system tables, with no configuration to create and no destination to choose. Query plans, session history, wait events, schema-change history, database and compute health, and Postgres logs land there, and records are retained for 7 days. Account admins and metastore admins can read the tables and manage access to them, and a user with access queries the telemetry the way they would any Unity Catalog table. See Lakebase telemetry in system tables.