Kafka Monthly Digest: February 2026
This 97th edition of the Kafka Monthly Digest covers what happened in the Apache Kafka community in February 2026.
Releases
There are 2 new releases (4.2.0 and 3.9.2) and 3 releases in progress (4.0.2, 4.1.2 and 4.3.0):
4.2.0
After 5 release candidates Christo Lolov released Kafka 4.2.0 on February 17. You can find the announcement on the Apache Kafka blog. You can also check the release notes and the release plan in the wiki. With 38 KIPs implemented (an all time record?) this release brings several major new features and many improvements across all components.
Kafka brokers and clients
Updates to the Kafka broker and clients include the following:
Share groups (queues) are now ready for production usage. (KIP-932) Improvements to share groups such as acquisition lock timeout renewal (KIP-1222), new lag metrics (KIP-1226) and configurations to enforce or not the max fetch record limit (KIP-1206) Consistency improvements to all command line tools. They now all support the --bootstrap-server and --command-config flags. (KIP-1147) Improvements and new features to many tools including kafka-consumer-perf-test.sh, kafka-producer-perf-test.sh, kafka-e2e-latency.sh and kafka-features.sh. (KIP-1052, KIP-1172, KIP-1192)
Kafka Connect
Updates to Kafka Connect include the following:
Support for external schemas in JSONConverter. (KIP-1054) New Allowlist connector client configuration override policy. This will be the new default policy from Kafka 5.0.0, but I recommend adopting it now if possible. (KIP-1188)
Kafka Streams
Updates to Kafka Streams include the following:
Support for dead letter queue (DLQ) in exception handlers. (KIP-1034) The new Streams rebalance protocol (based on KIP-848) is now production ready. The default protocol is still classic. (KIP-1071) Support for punctuation to be triggered at deterministic times. (KIP-1146)
3.9.2
PoAn Yang released Kafka 3.9.2 on February 21. This bug fix release fixes 35 issues. You can find the announcement on the blog and the release plan in the wiki. 4.0.2
Viktor Somogyi-Vass published 4.0.2 RC1 on February 19. The vote is currently on-going. You can find the release plan in the wiki.
4.1.2
Andrew Schofield published 4.1.2 RC1 on February 25. The vote is currently on-going. You can find the release plan in the wiki. 4.3.0
I volunteered to be the release manager for Kafka 4.3.0. The first milestone was KIP freeze on February 27. The next milestone is feature freeze on March 18. The release is planned for May. You can find the release plan in the wiki. Kafka Improvement Proposals
Last month, the community submitted 14 KIPs (KIP-1278 to KIP-1292, KIP-1281 was skipped). I'll highlight a few of them: KIP-1279: Cluster Mirroring: The KIP proposes bringing built-in cluster mirroring support in the brokers. This is an alternative to MirrorMaker but it would bring significant improvements including exact mirroring with matching offsets between the clusters, and it would simplify operations such as failover and failback.
KIP-1288: SSL Hot Reload for Kafka Clients: Kafka clients load their credentials at startup. For keys and certificates that are regularly renewed, clients must be restarted to pick up the new ones. This KIP introduces a mechanism for clients to reload their credentials at runtime easing the operational burden of managing secured environments.
KIP-1289: Support Hierarchical Transactional Acknowledgments for Share Groups: At the moment it's not possible to obtain exactly once semantic processing using share groups (queues). This KIP proposes new APIs to enable transactional acknowledgments. This would allow stream processing frameworks like Apache Flink and Apache Spark to provide exactly once semantics when consuming data from share groups.
KIP-1292: New MirrorMaker2 Connector for syncing consumer offsets: In MirrorMaker the MirrorCheckpointConnector handles the offset translation between the clusters. Because it's not possible to maintain a mapping for every single offset, the algorithm it currently uses ensures very precise translation for offsets near the log end offsets but that precision degrades exponentially the older offsets are. This KIP proposes a new algorithm with a constant degradation from the log start to end offsets so it could provide much more accurate translation for older offsets.
Community Releases
I selected releases of some open source community projects:
Sarama 1.47: Sarama is a pure Golang Kafka client. This release improves the Admin DescribeCluster() API, and the handling of broken connections.
Blogs
I selected some interesting blog articles that were published last month:
Disclaimer: Please note the content in this blog post has not been thoroughly reviewed by the Red Hat Developer editorial team. Any opinions expressed in this post are the author's own and do not necessarily reflect the policies or positions of Red Hat.