Release Notes 26.0.650

1. RELEASE INFO

2. CONTENT OF THE RELEASE

2.1 Highlights

• New External API endpoints of entity import/export in YAML format for improved support of agentic AI workflows, in Private Preview (for selected customers)

• Workflow Orchestration: Apache Airflow 3 upgrade in Private Preview (for selected customers)

• New feature Review Request in Private Preview (for selected customers)

• New STORAGE.CHANGE_TRACKING support in Databricks SQL

• New API_CALL support for Databricks API endpoints

2.2 Fixes

• Changing VIEW to TABLE leaves the VIEW SQL as hidden-in-UI load step (type of TRANSFORM) for the TABLE load.

• Load SQL generation fails for OPT_HISTORIZED* and OPT_INSERT* load options, when only either DV_LOAD_TIME or DV_UPDATE_TIME attribute types are set

• Deployment fails to number of retries exhausted error, when auto-promotion is on for downstream environments other than DEV

• Microsoft Entra ID service principal authentication support does not work with Azure SQL Database

• Entity description generates misplaced COMMENT clause in CREATE SEMANTIC VIEW

• Change in STORAGE.PARTITION_SCHEME does not result to entity re-create with ICEBERG_TABLEs

• Moving an entity with STORAGE.REQUIRE_PARTITION_FILTER = true to another schema fails

• Missing attribute type should still have default fallback with OPT_SCAN_LATEST_BY_KEY

3. NOTIFICATIONS

3.1 When the tenant has Apache Airflow 3 upgraded

When the tenant has Apache Airflow 3 upgraded for Workflow Orchestration, it starts from the clean state with its database (recommended by the open source project). The past executions and logs for Apache Airflow 2 can be accessed around 30 days by modifying the web browser URL from /dagger to /archiveddagger. The so-called archived Dagger ie. Airflow 2 is in read-only mode.

The main UIs of Apache Airflow 3