All Products
Search

This Product

    Realtime Compute for Apache Flink:October 13, 2025

    Document Center

    Realtime Compute for Apache Flink:October 13, 2025

    Last Updated:Oct 22, 2025

    This topic describes the major feature updates and bug fixes in the Realtime Compute for Apache Flink release on October 13, 2025.

    Important

    The version upgrade is gradually released to users. For more information, see the latest announcement in the Realtime Compute for Apache Flink console. You can use the new features in this version only after the upgrade is complete for your account. To apply for an expedited upgrade, submit a ticket.

    Overview

    A new version of Realtime Compute for Apache Flink was released on October 13, 2025. This release includes updates to the platform, engine, and connectors, along with performance optimizations and bug fixes.

    Engine

    The engine continues to evolve, enhancing support for complex data types, AI capabilities, and a wider range of data sources.

    Data ingestion

    Kafka source: The Canal-JSON format now supports MySQL_TYPE parsing.

    Connectors

    • Hologres connector: Supports reading specified binary log types.

    • StarRocks connector: The StarRocks connector is upgraded to align with StarRocks Connector for Flink release 1.2.11.

    • Lindorm connector: Dimension tables now support one-to-many joins.

    • MaxCompute connector: Supports custom time zone settings.

    • SLS connector: Added the processor parameter to pre-filter data from SLS before consumption, reducing costs and improving processing speed.

    AI capabilities

    • Vector search: Introduces the VECTOR_SEARCH function for real-time similarity searches using Milvus.

    • AI function: Adds more configurable parameters for greater flexibility.

    SQL and data management

    • JSON conversion: Flink SQL now supports converting an array in a JSON string to an ARRAY<STRING>.

    • StarRocks Catalog: You no longer need to explicitly pass a URL when using StarRocks Catalog.

    Performance optimizations

    • MySQL connector: Optimized predicate pushdown for boolean types.

    • Hologres connector: Added support for projection pushdown and data compression when consuming binary logs.

    • MaxCompute connector: Improved write performance for complex types like arrays into regular tables.

    Platform

    New capabilities

    • Workflows (public preview) is now available in US (Silicon Valley).

    • Hybrid billing is now available for YITIAN ARM workspaces in China (Ulanqab).

    Platform experience enhancements

    • Improved purchase experience

      • Removed CU limit: The 1000 CU limit for subscription workspaces has been removed for scale-ups, new purchases, and conversions from pay-as-you-go to subscription. This enables elastic scaling of large-scale jobs and offers greater flexibility for new purchases.

    • Enhanced observability

      • Unified GC metric display: Added support for garbage collection (GC) types such as ParallelGC and CMS. This enables unified monitoring and management, improving operational consistency and observability.

      • New busyTimePerSecond metric: This metric provides precise insights into job load to help you identify performance bottlenecks, allocate resources effectively, and ensure job stability.

    • Optimized permission model

      Folder permissions are now unified with job draft and data query permissions. Users can now create folders without permission errors, improving system clarity and user experience.

    • Improved materialized table feature

      Added support for upgrading the VVR version as you edit a materialized table. It also supports version comparison and automatic job deployment update, significantly improving usability and configuration consistency.

    • Flexible message delivery configuration

      Supports customizing the delivery scope for messages, for more granular control and cost optimization.

    Feature summary

    Feature

    Description

    References

    Hologres concurrent batch writes

    Enables concurrent writes to Hologres tables with a primary key after a reshuffle, to increase write throughput.

    Connector options (VVR 11 or later)

    Hologres binary log consumption optimization

    Supports projection pushdown and data compression when consuming binary logs to reduce network and computational overhead.

    Specify Hologres binary log type

    Allows users to specify the type of binary log to read, enabling more granular data sync.

    Kafka Canal-JSON MySQL_TYPE parsing

    The Canal-JSON format for the Kafka source now supports parsing native MySQL types.

    Kafka

    JSON String to ARRAY<STRING> conversion

    Flink SQL now supports the JSON_QUERY function that converts an array in a JSON string directly into an ARRAY<STRING>.

    JSON functions

    AI vector search

    Adds a new AI function for vector search.

    VECTOR_SEARCH

    MaxCompute time zone setting

    The MaxCompute connector now supports custom time zones for easier cross-time-zone data processing.

    MaxCompute

    SLS consumption processor

    Uses the processor parameter to pre-filter data from SLS before consumption, reducing the computational load on Flink.

    SLS

    Notable fixes

    This release resolves the following major bugs:

    Connector fixes

    • Fixed a deadlock issue that occurred when reading data with PostgreSQL CDC.

    • Fixed an issue where the Kafka connector would drop time zone information when converting timestamps that have time zone attributes.

    • Resolved a NullPointerException that occurred during joins with a Lindorm dimension table.

    • Fixed an issue where viewing or deleting a Paimon table after a parameter misconfiguration would result in a Could not find any factory for identifier 'last_not_null_value' error.

    SQL fixes

    • Fixed a format issue with the RETURNING ARRAY<STRING> clause in the JSON_QUERY function.