Skip to main content

​
Key Features

​
Added

  • Added Apache Doris as a destination.
  • Enhanced the Apache Kafka destination connector.
    • Schema registry is now optional, with JSON as the default registry-free message format.
    • Added flexible topic-routing options.
    • Added partitioning strategy options.
    • Added operational metadata to Kafka message headers.
  • Added destination schema management options during job creation. You can now choose to preserve the source schema layout in the destination or consolidate all tables into a single destination schema. Missing destination schemas are verified and automatically created at replication time.
  • Added support for running pipeline steps in parallel, reducing overall pipeline run time.
  • Added support for setting pipeline variables.
  • Added a live records-affected count for running tasks. The UI now displays a running total of records committed to the destination while a sync task is executing instead of reporting the total only after the task completes.
  • Added the SnapshotMode replicate option (Incremental / Full / Skip) to control how initial snapshots are performed.
  • Added the PreserveCharDatatype replicate option to maintain CHAR(n) source data types in the destination.
  • Added support for multiple change data capture (CDC) instances in Microsoft SQL Server CDC jobs, preserving change-capture continuity across schema changes.
  • Added support for incremental replication when you use SingleStore as a source.
  • Added support for relative record number (RRN) selection using the RRN(Table) function in DB2 for i CDC replication queries.
  • Added automatic trigger regeneration for SAP HANA CDC when a schema change is detected.
  • Added support for the CONCAT, CAST, FORMAT, and DATETIMEFROMPARTS functions in the CDC transformation library.

​
Changed

  • Expanded LocalDateTime support to all destinations and CDC sources. New jobs now automatically enable UseLocalDateTime when both the source and destination support it, preserving timestamp values without time-zone conversion.
  • Improved bulk-loading performance for Databricks, Snowflake, Azure Synapse, Amazon Redshift, Google BigQuery, CSV, Avro, and Parquet destinations, especially for wide tables with many columns.
  • Improved PostgreSQL destination performance by using the COPY statement for bulk inserts.
  • Improved snapshot merge behavior to use transaction boundaries, avoiding timeouts during large initial loads.
  • Consolidated the Replicate Engine status tables into a single unified status table. The UI now notifies you when your existing status table can be upgraded, and a new PARTIAL_SUCCESS task status was added for more accurate job reporting.
  • Enhanced replication queries can now execute in a separate Replicate Agent child process, improving isolation and stability of the application. This feature is not available when the application database is Derby or H2.
  • Improved the Add Tasks dialog. The task search now supports filtering out schemas, and Microsoft SQL Server CDC jobs only list tables with CDC or change tracking enabled.
  • Updated the default SyncIdSourceColumn value for new Oracle and DB2 for i CDC jobs to OracleRowId and Db2Rrn, respectively. For DB2 for i tables without a primary key, now detects a unique index on NOT NULL columns and uses it as the merge key instead of falling back to the RRN. also logs a warning when an RRN is used or when a unique index is skipped because it contains nullable columns.
  • Improved incremental replication for sources that can return the incremental check column in sorted order. For these sources, automatically retrieves the column’s minimum value to determine the starting point (so you no longer need to configure a start value), reads rows in order, and commits progress as it goes. This behavior allows an interrupted run to resume from the last committed value instead of starting over.
  • Improved Git version control with proxy support, a redesigned initialization and setup flow, and elimination of timestamp-only .manifest changes from workspace commits.
  • Updated the Role Policy UI to warn you about related permissions that are required when granting Create Pipeline permission.
  • Added startup blocking and email notification when application-database table initialization fails.
  • Updated the API Access permission to be enabled automatically when you generate an auth token from User Details.
  • Refreshed the Dashboard, Jobs, Connections, Transformations, Logs, and detail pages with updated shared UI components for a more consistent experience.
  • Migrated the documentation to a new documentation platform. In-application help links now point to the new site.

​
Fixed

  • Fixed an issue in the Admin API V2 where the update-connection endpoint returned a JSON array instead of a JSON object.
  • Fixed an issue in the code-view editor for the API connector where newly added attributes were not saved and were lost after reloading.
  • Fixed an issue where the last-visited workspace was incorrectly displayed at login.
  • Fixed an issue where the Log File Verbosity drop-down label did not update until you clicked Save.
  • Fixed an issue where the source connector icon was missing in the Name column on the Job History tab.
  • Fixed an issue where filters did not work as expected on the Activity page.
  • Fixed an issue where the Git Version Control help button returned a 404 error.
  • Fixed several minor UI issues.