Apache Kafka 4.2.0 Release Announcement

We are proud to announce the release of Apache Kafka® 4.2. This release contains many new features and improvements, and this blog post will highlight some of the more prominent ones. For a full list of changes, be sure to check the release notes.

Kafka Queues (Share Groups) is now production-ready with new features like the RENEW acknowledgement type for extended processing times, adaptive batching for share coordinators, soft and strict enforcements of quantity of fetched records, and comprehensive lag metrics.

Kafka Streams brings the server-side rebalance protocol to GA with a limited feature set, adds dead letter queue support in exception handlers, introduces anchored wall-clock punctuation for deterministic scheduling, and gives users full control over whether to send a leave group request on closing.

This release also delivers significant improvements to consistency and observability: CLI tools now feature standardized arguments like --bootstrap-server across all tools, metric naming has been corrected to follow the kafka.COMPONENT convention, and new idle ratio metrics provide better visibility into controller and MetadataLoader performance.

Security is enhanced with a new allowlist connector client configuration override policy, while thread-safety improvements to RecordHeader eliminate concurrency risks.

Additional highlights include support for Java 25,  external schema support in JsonConverter for reduced message sizes, dynamic configuration for remote log manager thread pools, adaptive batching in group coordinators, and rack ID exposure in the Admin API for consumer and share group members.

See the Upgrading to 4.2 section in the documentation for the list of notable changes and detailed upgrade steps. Listen as Sandon Jacobs (Senior Developer Advocate, Confluent) highlights the key features in Apache Kafka 4.2 in this video.

Deprecation Notices

• KIP-1136: Make ConsumerGroupMetadata an interface: Deprecates the constructors of ConsumerGroupMetadata. Marked for removal in Apache Kafka 5.0.

• KIP-1193: Deprecate MX4j support: Adds various deprecation warnings for MX4j. Marked for removal in Apache Kafka 5.0.

• KIP-1195: deprecate and remove org.apache.kafka.streams.errors.BrokerNotFoundException: Deprecates BrokerNotFoundException. Marked for removal in Apache Kafka 5.0.

Kafka Broker, Controller, Producer, Consumer and Admin Client

• KIP-932: Queues for Kafka: Introduces share groups, a new cooperative consumption model where multiple consumers can concurrently process records from the same partitions with individual acknowledgment and delivery counting - enabling queue-like use cases without strict partition-to-consumer assignment.

• KIP-1052: Enable warmup in producer performance test: Adds an optional --warmup-records argument to the producer performance test, separating startup measurements from steady-state statistics for cleaner performance analysis.

• KIP-1100: Rename org.apache.kafka.server:type=AssignmentsManager and org.apache.kafka.storage.internals.log.RemoteStorageThreadPool metrics: Fixes inconsistent metric naming by deprecating metrics accidentally changed to org.apache.kafka.COMPONENT format and introducing new metrics using the correct kafka.COMPONENT convention.

• KIP-1147: Improve consistency of command-line arguments: Standardizes command-line tool arguments by introducing consistent options like --bootstrap-server and --command-config across all tools, deprecating inconsistent legacy options.

• KIP-1157: Enforce KafkaPrincipalSerde Implementation for KafkaPrincipalBuilder: Makes KafkaPrincipalBuilder extend KafkaPrincipalSerde to enforce serialization/deserialization support at compile-time rather than failing at runtime during KRaft broker-controller communication.

• KIP-1160: Enable returning supported features from a specific broker: Adds an optional --node-id argument to describeFeatures, allowing users to query supported features from a specific node to address version inconsistencies across nodes.

• KIP-1161: Unifying LIST-Type Configuration Validation and Default Values: Standardizes validation for comma-separated list configurations by rejecting null/empty values, ignoring duplicates, converting string configs to proper LIST types, and throwing ConfigExceptions during parsing rather than at runtime.

• KIP-1172: Improve EndToEndLatency Tool with argument parser and message key/header support: Improves the EndToEndLatency tool with robust named-argument parsing, new optional parameters for message keys and headers, and renamed arguments aligned with Kafka conventions.

• KIP-1175: Fix the typo PARTITIONER_ADPATIVE_PARTITIONING_ENABLE in ProducerConfig: Fixes a misspelling by introducing the correctly spelled PARTITIONER_ADAPTIVE_PARTITIONING_ENABLE_CONFIG constant while deprecating the misspelled version.

• KIP-1179: Introduce remote.log.manager.follower.thread.pool.size config: Introduces a new dynamic configuration remote.log.manager.follower.thread.pool.size with proper naming conventions and dynamic configuration support for the follower partition thread pool.

• KIP-1180: Add generic feature level metrics: Adds new metrics displaying finalized, minimum supported, and maximum supported feature levels for each production feature, improving visibility into upgrade/downgrade scenarios.

• KIP-1186: Update AddRaftVoterRequest RPC to support auto-join: Adds a boolean AckWhenCommitted flag to AddRaftVoterRequest that allows immediate response after locally writing the new voter set, preventing availability issues during auto-joining controller operations.

• KIP-1190: Add a metric for controller thread idleness: Adds a new AvgIdleRatio metric measuring the proportion of time the controller thread spends idle versus actively processing events, improving performance visibility.

• KIP-1192: Add include argument to ConsumerPerformance tool: Adds an --include argument to ConsumerPerformance for regex-based topic filtering, enabling multi-topic performance testing.

• KIP-1197: Introduce new method to improve the TopicBasedRemoteLogMetadataManager’s initialization: Fixes TopicBasedRemoteLogMetadataManager initialization failures by introducing a BrokerReadyCallback interface that delays initialization until the broker is fully ready.

• KIP-1205: Improve RecordHeader to be Thread-Safe: Addresses thread-safety issues in RecordHeader by implementing double-checked locking with volatile fields, eliminating NullPointerException risks during concurrent access with negligible overhead.

• KIP-1206: Strict max fetch records in share fetch: Introduces a new ShareAcquireMode configuration for shared fetch operations, offering “batch_optimized” (soft limit) and “record_limit” (strict enforcement) modes for different processing scenarios.

• KIP-1207: Fix anomaly of JMX metrics RequestHandlerAvgIdlePercent in kraft combined mode: Fixes the RequestHandlerAvgIdlePercent metric in KRaft combined mode by normalizing by combined thread count and introducing separate broker and controller metrics for per-pool visibility.

• KIP-1217: Include push interval in ClientTelemetryReceiver context: Addresses stale client telemetry metrics by introducing new interfaces that include push interval information, enabling proper metric lifecycle management.

• KIP-1222: Acquisition lock timeout renewal in share consumer explicit mode: Adds a new RENEW acknowledgement type for share consumers, allowing applications to extend acquisition lock timeouts on records requiring longer processing times.

• KIP-1224: Adaptive append.linger.ms for the group coordinator and share coordinator: Introduces adaptive batching mode for group and share coordinators that automatically adjusts batch linger times based on workload, eliminating the 5ms latency floor without manual tuning.

• KIP-1226: Introducing Share Partition Lag Persistence and Retrieval: Adds share partition lag metrics for share groups, enabling operators to monitor consumption progress, detect imbalances, and support future autoscaling capabilities.

• KIP-1227: Expose Rack ID in MemberDescription and ShareMemberDescription: Exposes rack ID information for consumer and share group members in the Admin API by adding a rackId field to member description classes.

• KIP-1228: Add Transaction Version to WriteTxnMarkersRequest: Adds a TransactionVersion field to WriteTxnMarkersRequest, enabling stricter epoch validation for Transaction Version 2 markers and strengthening exactly-once semantics guarantees.

• KIP-1229: Add Idle Thread Ratio Metric to MetadataLoader: Adds an AvgIdleRatio metric to the MetadataLoader component in KRaft clusters, improving visibility into event queue processing efficiency.

Kafka Streams

• KIP-1034: Dead letter queue in Kafka Streams: Adds dead letter queue (DLQ) support to Kafka Streams exception handlers by introducing a new Response class with DLQ records, new handleError() methods, and raw source record bytes in error contexts.

• KIP-1071: Streams Rebalance Protocol: Introduces a new server-side group management protocol for Kafka Streams, enabling broker-side task assignment, centralized topology metadata storage, and improved observability through dedicated RPCs and admin tools.

• KIP-1146: Anchored punctuation: Introduces anchored wall-clock punctuation for Kafka Streams by adding an optional startTime parameter to schedule(), enabling callbacks at fixed, deterministic times (e.g., exactly at the start of every hour).

• KIP-1153: Refactor Kafka Streams CloseOptions to Fluent API Style: Gives users explicit control over whether KafkaStreams sends a leave-group request on shutdown via a new GroupMembershipOperation enum, wrapped in a fluent-style CloseOptions class that replaces the deprecated boolean-based API.

• KIP-1216: Add rebalance listener metrics for Kafka Streams: Adds thread-level latency metrics for tasks-revoked, tasks-assigned, and tasks-lost rebalance callbacks in Kafka Streams, restoring observability after the move to the dedicated streams rebalance protocol.

• KIP-1221: Add application-id tag to Kafka Streams state metric: Adds an application-id tag to the Kafka Streams client-state JMX metric, enabling operators to group multiple instances belonging to the same logical application.

• KIP-1230: Add config for file system permissions: Adds an opt-in allow.os.group.write.access configuration for Kafka Streams, allowing users to grant write access to their OS user-group for the local state directory.

Kafka Connect

• KIP-1054: Support external schemas in JSONConverter: Adds an optional schema.content configuration to JsonConverter, allowing schemas to be specified externally rather than embedded in every JSON message - reducing message size and simplifying plain JSON integration.

• KIP-1120: AppInfo metrics don’t contain the client-id: Adds a client-id tag to AppInfo metrics for Kafka Worker and MirrorMaker 2 clients, improving monitoring and debugging consistency with other Kafka clients.

• KIP-1188: New ConnectorClientConfigOverridePolicy with allowlist of configurations: Introduces a new “Allowlist” connector client configuration override policy to address security vulnerabilities, allowing administrators to explicitly specify which client configurations connectors can override.

Summary

Ready to get started with Apache Kafka 4.2.0? Check out all the details in the upgrade notes and the release notes, and download Apache Kafka 4.2.0.

This was a community effort, so thank you to everyone who contributed to this release, including all our users and our 155 contributors: Abhi Tiwari, Abhijeet Kumar, Abhinav Dixit, Abhiram98, Alex, Alieh Saeedi, ally heev, Alyssa Huang, Andrew J Schofield, Anton Vasanth, Apoorv Mittal, Arpit Goyal, Artem Livshits, Bill Bejeck, Bolin Lin, Bruno Cadonna, Calvin Liu, Chang-Chi Hsu, Chang-Yu Huang, Chia-Ping Tsai, Chih-Yuan Chien, Chirag Wadhwa, Chris Egerton, Christo Lolov, Chuckame, Clemens Hutter, Colin Patrick McCabe, d00791190, Dave Troiano, David Arthur, David Jacot, Deep Golani, Dejan Stojadinović, devtrace404, Dmitry Werner, Dongnuo Lyu, Donny Nadolny, Eduwer Camacaro, Elizabeth Bennett, EME, Eric Chang, Erik Anderson, Evan Zhou, Evgeniy Kuvardin, farzan ghalami, Fatih, Federico Valeri, Gantigmaa Selenge, Gasparina Damien, Gaurav Narula, Genseric Ghiro, George Wu, Greg Harris, Harish Vishwanath, Herman Kolstad Jakobsen, Hong-Yi Chen, Ismael Juma, Izzy Harker, Jared Harley, Jhen-Yung Hsu, Jian, Jim Galasyn, Jimmy Wang, Jing-Jia Hung, Jinhe Zhang, Joel Hamill, Jonah Hooper, Josep Prat, José Armando García Sancio, Juha Mynttinen, Jun Rao, Justine Olshan, k-apol, Kamal Chandraprakash, Kaushik Raina, keemsisi, Ken Huang, Kevin Wu, Kirk True, knoxy5467, KTKTK-HZ, Kuan-Po Tseng, Lan Ding, Levani Kokhreidze, Liam Clarke-Hutchinson, Lianet Magrans, Linsiyuan9, Logan Zhu, lorcan, Lord of Abyss, Lucas Brutschy, Lucy Liu, Luke Chen, Mahsa Seifikar, majialong, Manikumar Reddy, Maros Orsak, Masahiro Mori, Mason Chen, Matt Welch, Matthias J. Sax, Michael Knox, Michael Morris, Mickael Maison, Ming-Yen Chung, NeatGuyCoding, Nick Guo, NICOLAS GUYOMAR, Nikita Shupletsov, Now, Okada Haruki, Omnia Ibrahim, Otmar Ertl, OuO, Paolo Patierno, Patrik Nagy, Pawel Szymczyk, PoAn Yang, Ken Huang, Priyanka K U, Rajani K, Rajini Sivaram, Ram, Ritika Reddy, Robert Young, Ryan Dielhenn, S.Y. Wang, samarth-ksolves, Sanskar Jhajharia, Satish Duggana, Sean Quah, Sebastien Viale, Shang-Hao Yang, Shashank, Shivsundar R, Siyang He, Sophie Blee-Goldman, Stig Døssing, stroller, Sushant Mahajan, TaiJuWu, TengYao Chi, Tsung-Han Ho (Miles Ho), Ubuntu, Uladzislau Blok, Vincent PÉRICART, Xiao Yang, xijiu, Xuan-Zhang Gong, yangxuze, Yeikel Santana, Yu-Syuan Jheng, YuChia Ma, Yunchi Pang, Yung

This blog post was originally published by Christo Lolov on The Apache Software Foundation blog.

Apache®, Apache Kafka®, and  Kafka® are either registered trademarks or trademarks of the Apache Software Foundation in the United States and/or other countries. No endorsement by the Apache Software Foundation is implied by using these marks. All other trademarks are the property of their respective owners.