Skip to main content
Version: Next

1.44.0 - 2026-02-16

Added​

  • Client: Add JWT authentication support #4313 @jakub-moravec Add JWT authenticator for Java and Python clients, enabling token-based authentication without requiring a custom authenticator implementation.
  • Spark: Extract input symlinks from DataSourceRDD #4283 @kchledowski Enable extraction of input dataset symlinks from DataSourceRDD, providing richer lineage information for RDD-based Iceberg operations.

Changed​

  • Spark: Disable column-level lineage for LogicalRDD plans #4329 @kchledowski Disable column-level lineage extraction for LogicalRDD plans to prevent incorrect lineage caused by lost schema and transformation context.
  • Spark: Disable input schema extraction for LogicalRDD and add schema extraction for Iceberg DataSourceRDD #4331 @kchledowski Disable unreliable input schema extraction from LogicalRDD and instead extract schemas from Iceberg table metadata when reading via DataSourceRDD.

Fixed​

  • dbt: Align schema definition for dbt facets #4285 @LegendPawel-Marut Align schema definitions for dbt-run-run-facet and dbt-version-run-facet to fix validation inconsistencies.
  • dbt: Better error handling for finding the directory of profiles.yml #4320 @ah12068 Handle missing profiles_dir key in run_results.json gracefully, falling back to default profile directory resolution.
  • dbt: Fix --target-path CLI flag being ignored #4298 @gaurav-atlan Fix the --target-path CLI argument not being parsed and passed to artifact processors, causing the default target path to always be used.
  • Flink: Use Flink 2.x-only class for version detection #4312 @mobuchowski Fix false Flink 2.x detection when modern V2-based connectors are used with Flink 1.x by using JobStatusChangedListenerFactory for version detection.
  • Java: Send Content-Encoding header for compressed requests #4282 @Lukas-Riedel Send the Content-Encoding header when request body compression is enabled in the Java client, consistent with the Python client behavior.
  • Spark: Early return in DatabricksEventFilter for non-Databricks platforms #4315 @mobuchowski Add fast environment detection check to skip Databricks-specific event filtering on non-Databricks platforms, reducing overhead.
  • Spark: Handle null Glue ARN in IcebergHandler #4311 @mobuchowski Fix NullPointerException when processing Iceberg datasets with AWS Glue catalog by safely handling null Glue ARN values.
  • Spark: Handle ResolvedIdentifier in DropTableVisitor #4316 @mobuchowski Fix ClassCastException on DROP TABLE commands in Databricks Runtime 14.2+ by handling ResolvedIdentifier alongside ResolvedTable.
  • Python: Remove unneeded build runtime dependency #4344 @mobuchowski Remove the build package from runtime dependencies as it is only needed at build time and is already handled by the build system configuration.
  • Spark: Resolve parent job name in ParentRunFacet for AWS Glue #4340 @tstrilka Fix incorrect parent job name in ParentRunFacet for child events (SQL_JOB, RDD_JOB) on AWS Glue, where the raw spark.app.name was used instead of the resolved application name from platform-specific name resolvers.
  • Spark: Skip Iceberg RDD integration tests on Java 8 with Spark 3.5 @mobuchowski Skip testAppendWithRDDTransformations and testAppendWithRDDProcessing on Java 8 + Spark 3.5 where the Iceberg vendor module is not compiled due to Iceberg 1.7 requiring Java 11+.