v0.145.0
Summary
Job definitions now support trigger configuration and MLflow artifact location for AI runtime tasks. Pipeline connector options and GCP endpoint settings have been expanded with new configuration fields.
Summary generated by brickster.ai. For the full changelog and any code/binary attachments, follow the GitHub link above.
More from databricks/databricks-sdk-java
The SDK adds comprehensive IAM v2 API methods to both account and workspace clients for managing groups, users, service principals, and workspace assignments through create, read, list, and delete operations. The `newCluster` field in `JobCluster` is no longer required, making job cluster configuration more flexible.
The SDK adds support for user token forwarding in apps, custom Docker images in AI runtime jobs, and GPU_XLARGE_8 serving options, plus a secret value retrieval field. ServicePrincipal and User objects now require additional fields (accountSpStatus, displayName for principals; accountUserStatus, fullName, username for users), which is a breaking change for code managing IAM principals or users.
The release adds `inputColumn` and `outputColumn` fields to the Transformer class and enables NetSuite as a connection type. Multiple breaking changes remove `includeBrowse` request parameters and `browseOnly` fields from MCP, ModelProvider, and Model service catalog APIs.
Fixed a critical bug where file uploads with streaming bodies silently sent empty content on retry, with the SDK now surfacing the original error instead so callers can retry with a fresh stream. Added new fields and enum values to billing and catalog APIs, including principalOverrides, scopeType, resourceType, and grant fields to expand billing configuration and policy management capabilities.
* Bump Jackson and Apache Commons Configuration dependencies to address security vulnerabilities.
This release adds Kinesis stream configuration support for ML models and a mode field for pipeline updates. Multiple breaking changes remove direct credential fields from AWS, Azure, and Microsoft model provider configurations and convert state fields to String type in bundle deployment APIs.