This 98th edition of the Kafka Monthly Digest covers what happened in the Apache Kafka community in March 2026.
Releases
There are 2 new releases (4.0.2 and 4.1.2) and 2 releases are in progress (4.3.0 and 4.2.1):
4.0.2
Viktor Somogyi-Vass released 4.0.2 on March 16. This bug fix release fixes 43 issues, including several blocker Jiras. It also updates dependencies (log4j, lz4, jetty) to resolve CVEs. You can find the announcement on the blog and the release plan in the wiki.
4.1.2
Andrew Schofield released 4.1.2 on March 17. This bug fix release fixes 31 issues. Like 4.0.2, it updates dependencies (log4j, lz4, jetty) to resolve CVEs. You can find the announcement on the blog and the release plan in the wiki.
4.3.0
The release process for Kafka 4.3.0 continued. Feature freeze happened on March 18. The next milestone is code freeze on April 8. The release is planned for May. You can find the release plan in the wiki.
4.2.1
On March 25, PoAn Yang volunteered to run the 4.2.1 release. You can find the release plan in the wiki.
Kafka Improvement Proposals
- KIP-1295: Dead letter queue support for global store/KTable: This KIP extends KIP-1270 which enabled exceptions raised during GlobalKTable record processing to be handled by a ProcessingExceptionHandler. With this new KIP, records which failed during GlobalKTable processing could be sent to a dead letter queue.
- KIP-1303: Deprioritize Tiered Storage Followers In Leader Election: From Kafka 4.3.0 (KIP-1023), when adding a new replica to a tiered topic, users can start replication from the earliest pending upload offset, meaning the new replica only copies data that the current leader has not yet sent to remote storage and ignores data on the remote storage. In practice this allows new replicas to become in-sync much faster. The tradeoff is that the new replica hosts a lot less local data so if it becomes the leader it's more likely to have to fetch segments from the remote storage which can degrade performance. KIP-1303 proposes a mechanism to prefer replicas with earlier local-log-start-offset when running leader election for topics with tiered storage enabled.
- KIP-1307: SerDe and Interceptor Metrics: Overall Kafka clients are well instrumented and provide many metrics to understand their behavior and assess their performance. However one weak spot is in the execution of SerDes and interceptors. This KIP proposes adding new metrics that track the behavior or Serializers, Deserializers, and Producer and Consumer interceptors with the time spent in each method and their error rate.
Community Releases
I selected releases of some open source community projects:
- strimzi-kafka-operator 0.51: Strimzi is a Kubernetes Operator for running Kafka. This release adds support for Kafka 4.2.0. It also enables configuring connection settings per listener, and server-side applies is now enabled by default for PersistentVolumeClaim, ServiceAccount, Service, Ingress and ConfigMap resources.
- kroxylicious 0.19.0: Kroxylicious is an open source pluggable framework for writing network proxies that understand the Apache Kafka protocol. This release also adds support for Kafka 4.2.0. The authorizer filter has been updated to behave like Kafka when checking transactional Ids. Users can now configure idle timeouts on authenticated and unauthenticated connections.
Blogs
I selected some interesting blog articles that were published last month:
