Folge Databricks, um die Release Notes in deinen Feed zu holen.
OpenAI GPT OSS 120B und GPT OSS 20B werden nun von On-demand Provisioned Throughput Foundation Model APIs unterstützt.
The Databricks-hosted foundation models, OpenAI GPT OSS 120B and GPT OSS 20B are now supported on On-demand provisioned throughput Foundation Model APIs.
Serverless GPU Compute unterstützt jetzt Multinode-Workloads.
Serverless GPU compute now supports multinode workloads. See AI Runtime.
Für Workloads auf Serverless GPU Compute lassen sich nun Jobs erstellen und zeitlich planen.
You can now create and schedule jobs for your Serverless GPU Compute workloads. See Schedule a notebook.
Nutzer können den Databricks Assistant über eigene Anweisungen zu Präferenzen, Coding-Konventionen und Antwortrichtlinien anpassen, und Admins können workspaceweite Anweisungen mit höherer Priorität festlegen.
You can now tailor how Databricks Assistant responds by adding custom user instructions. Guide the Assistant with preferences, coding conventions, and response guidelines.
Admins can also add workspace-wide instructions, which take priority over user instructions.
Instructions apply across Assistant experiences like Inline Assistant, General Chat, Suggest Fix, and Edit mode. See Customize Genie Code with custom instructions.
Single-Node-Compute mit Standard Access Mode ist allgemein verfügbar und erlaubt mehreren Nutzern, eine Single-Node-Ressource mit voller Nutzerisolation gemeinsam zu verwenden.
Single node compute resources with standard access mode is now generally available. This configuration allows multiple users to share a single-node compute resource with full user isolation. Single-node compute is useful for small jobs or non-distributed workloads.
See Compute configuration reference.
Beim Ersetzen einer Tabelle bleibt die bestehende Column Mask einer Spalte mit gleichem Namen erhalten, auch wenn keine Maske angegeben wird, während zuvor alle bisherigen Masken entfernt wurden.
If a column in the new table matches a column name from the original table, its existing column mask is now retained, even if no mask is specified. This change prevents accidental removal of column-level security policies during table replacement. Previously, replacing a table dropped all existing column masks, and only newly defined masks were applied.
This change affects SQL commands ([CREATE OR] REPLACE TABLE), DataFrame APIs (saveAsTable, replace, createOrReplace), and other similar update table operations.
See Manually apply row filters and column masks.
In Unity Catalog lassen sich Self-Service-Zugriffsanfragen aktivieren, die an konfigurierte Ziele wie E-Mail, Slack, Microsoft Teams oder ein internes Zugriffsverwaltungssystem gesendet werden.
You can now enable self-service access requests in Unity Catalog by configuring access request destinations on securable objects.
Users can request access to Unity Catalog objects that they discover. These requests are sent to configured destinations, such as emails, Slack or Microsoft Teams channels, or they can be redirected to an internal access management system.
You can also enable default email destinations so requests automatically go to the catalog or object owner’s email if no other destination is set. This ensures that access requests are delivered even when no destination is manually configured for an object.
See Manage access requests.
Account-Admins können per neuer Einstellung in der Account Console Legacy-Funktionen wie DBFS root, Hive metastore, No-isolation shared compute und Databricks Runtime vor 13.3 LTS in neuen Workspaces deaktivieren.
A new account console setting allows account admins to disable certain legacy features on new workspaces created in their account. If set, new workspaces will not include DBFS root and mounts, the Hive metastore, no-isolation shared compute, or Databricks Runtime versions prior to 13.3 LTS. If needed, workspace admins can still enable these legacy features from their workspace settings.
See Disable access to legacy features in new workspaces.
Die Vorschau von Lakebase: Managed Postgres OLTP Database ist standardmäßig aktiviert, sodass Nutzer Datenbankinstanzen ohne vorherige Admin-Freigabe erstellen können, während Admins sie deaktivieren können.
The Lakebase: Managed Postgres OLTP Database preview is now enabled by default. Users can start creating Lakebase database instances without needing workspace admins to enable this preview first. Admins can disable the preview if needed.
See Manage projects and Manage Databricks previews.
Serverless Workspaces mit vorkonfiguriertem Serverless Compute und Standardspeicher sind in der Public Preview verfügbar und können von Account-Admins in der Account Console aktiviert werden.
Serverless workspaces come pre-configured with serverless compute and default storage to provide a completely serverless experience. Account admins can enable the serverless workspaces Public Preview in the account console. See Create a serverless workspace.
Jobs im Continuous Mode können nun einzelne fehlgeschlagene Tasks erneut ausführen.
Jobs that are set to run in continuous mode now have the option to retry individual tasks on task failure.
See Run jobs continuously.
Der vollständig verwaltete ServiceNow-Ingestion-Connector in Lakeflow Connect ist allgemein verfügbar.
The fully-managed ServiceNow ingestion connector in Lakeflow Connect is now generally available. See Configure ServiceNow for Databricks ingestion.
Unity-Catalog-External-Tables auf Basis von Delta Lake lassen sich nun auch aus externen Clients und Systemen wie Apache Spark erstellen.
You can now create Unity Catalog external tables backed by Delta Lake from external clients and systems, such as Apache Spark.
See Access Databricks tables from Delta clients.
Mit Path Credential Vending lassen sich kurzlebige Zugangsdaten für External Locations im Unity-Catalog-Metastore an externe Systeme vergeben.
You can now use path credential vending to grant short-lived credentials to external locations in your Unity Catalog metastore. See Unity Catalog credential vending for external system access.
OpenAI GPT OSS 120B und GPT OSS 20B sind nun für AI Functions wie ai_query() optimiert, ermöglichen Batch Inference und unterstützen die Angabe von Structured Outputs.
The Databricks-hosted foundation models, OpenAI GPT OSS 120B and GPT OSS 20B are now optimized for AI Functions, which means you can perform batch inference using these models and AI Functions like ai_query(). Inference workloads that use these models also now support specifying structured outputs.
Serverless Compute für Notebooks, Workflows und Spark Declarative Pipelines auf Lakeflow ist nun in der Region Asia Pacific (Jakarta) ap-southeast-3 verfügbar.
Serverless compute for notebooks, workflows, and Spark Declarative Pipelines on Lakeflow is now available in the Asia Pacific (Jakarta) region (ap-southeast-3). See Features with limited regional availability.
Der Databricks ODBC Driver 2.9.2 nutzt den Prozessnamen als Standard-UserAgentEntry, unterstützt die Domains cloud.databricks.us und cloud.databricks.mil, timestamp_ntz- und VOID-Spalten, CRL-Cache unter Windows, OAuth-Token-Exchange für abweichende IDPs sowie Windows Server 2025, behebt ein Speicherleck, aktualisiert OpenSSL, libcURL und Expat und unterstützt Databricks Runtime 10.4 LTS nicht mehr.
The Databricks ODBC Driver version 2.9.2 is now available for download from the ODBC driver download page.
This release includes the following fixes and new features:
cloud.databricks.us and cloud.databricks.mil.timestamp_ntz columns across multiple data source functions including SQLGetTypeInfo, SQLColumns, and SQLColAttribute.UseSystemTruststore is enabled.VOID type column support, so that VOID columns are now correctly listed in SQLGetColumns.This release includes upgrades to several third-party libraries:
This release includes the following behavior changes:
Die Authentifizierung in Power BI Desktop ist jetzt mit Machine-to-Machine-(M2M)-OAuth-Zugangsdaten möglich, und Databricks empfiehlt Nutzern von Personal Access Tokens für Service Principals den Wechsel zur neuen Option „Client credentials“.
You can now authenticate into Power BI Desktop using Machine-to-Machine (M2M) OAuth credentials. If you are currently using personal access tokens on behalf of service principals, Databricks recommends switching to the new Client credentials authentication option.
See Configure service principals on Databricks for Power BI.
Mosaic AI Model Serving unterstützt jetzt OpenAIs GPT OSS 120B und GPT OSS 20B als von Databricks gehostete Foundation Models, nutzbar über Foundation Model APIs pay-per-token.
Mosaic AI Model Serving now supports OpenAI's GPT OSS 120B and GPT OSS 20B as Databricks-hosted foundation models. You can access these models using Foundation Model APIs pay-per-token.
AWS PrivateLink zu Ressourcen in der eigenen VPC über einen Network Load Balancer (NLB) lässt sich von Serverless Compute aus jetzt allgemein verfügbar aktivieren.
You can now enable AWS PrivateLink to resources in your VPC (virtual private cloud) through an NLB (network load balancer) from serverless compute. See Configure private connectivity to resources in your VPC.