Releases / Upgrades / Change logs

Apache Kafka 2.4

• Kafka Core
  • Allow consumers to fetch from closest replica (before this all reads and writes happened on the leader)
  • Implement admin API for replica reassignment
  • Sticky partitioner
  • Admin API for deleting consumers offset

https://www.confluent.io/blog/apache-kafka-2-4-latest-version-updates

Apache Kafka 4.0

Apache Kafka 4.0 Release: Default KRaft, Queues, Faster Rebalances

• Apache Kafka 4.0 | ZooKeeper out, KRaft in! Scalable, Resilient, High Performance, more Cloud-Native - YouTube
• Kafka Broker, Controller, Producer, Consumer and Admin Client
  • ZooKeeper is gone
  • KIP-848: The Next Generation of the Consumer Rebalance Protocol
  • KIP-890: Transactions Server-Side Defense
  • KIP-932: Queues for Kafka (Early Access)
    • Queues for Kafka Explained (KIP-932) - YouTube
    • Share Partition, Share Consumer
    • Cooperative consumption model
  • KIP-1106: Add duration based offset reset option for consumer clients
  • KIP-1043: Administration of groups
• Kafka Streams
  • KIP-1104: Allow Foreign Key Extraction from Both Key and Value in KTable Joins
  • KIP-1112: Allow custom processor wrapping
• Kafka Connect
  • KIP-1074: Allow the replication of user internal topics
    • Previously, MirrorMaker 2 automatically excluded topics whose names ended with .internal or -internal, incorrectly classifying them as internal topics. This behavior prevented legitimate business topics from being replicated unless users implemented a custom replication policy. This KIP introduces a configurable option that allows users to replicate such topics without requiring custom code.

Apache Kafka 4.1 | Enhanced Stability, New OAuth Support, Scalable Queue...

Others

Top Trends for Data Streaming with Apache Kafka and Flink in 2026 - Kai Waehner

1. Proven platforms gain ground as the data streaming ecosystem consolidates
2. Diskless Kafka and Apache Iceberg create a new storage foundation
3. Real-time analytics becomes part of the streaming layer
4. Enterprises demand SLAs with zero data loss and seamless failover
5. Regional cloud deployments are driven by compliance and sovereignty
6. Streaming powers Agentic AI with context and real-time model inference

Data streaming experts should focus on helping their organizations:

• Unify streaming and batch analytics in a single architecture
• Consolidate pipelines, reduce operational burden, and increase platform value
• Deliver regional compliance through sovereign deployment options
• Guarantee zero data loss and rapid failover across regions where needed
• Power AI systems with contextual, real-time event data