Apache Kafka 3.3 Release Highlights New KRaft Mode and Multiple KIP Enhancements
The Apache Kafka 3.3 release introduces production‑ready KRaft mode and a series of KIP enhancements across brokers, streams, and connect, detailing new metrics, API changes, and operational improvements while encouraging users to review the full release notes and try the updated platform.
We are proud to announce the release of Apache Kafka® 3.3, which brings a host of new features and improvements to the platform.
The new KRaft community mode, developed over several years, improves scalability and resilience and eliminates the need for an external Apache ZooKeeper™ cluster. KIP‑833 marks KRaft as production‑ready for new clusters.
Key broker‑related KIPs include: KIP‑833 (production‑ready KRaft), KIP‑778 (upgrade path for KRaft clusters), KIP‑841 (protected replicas in ISR), KIP‑836 (expose quorum metadata), KIP‑835 (monitor KRaft controller quorum health), KIP‑859 (metadata log error metrics), KIP‑794 (strict sticky partitioning), KIP‑373 (delegation tokens for other users), KIP‑831 (log recovery progress metrics), KIP‑709 (multi‑group OffsetFetch RPC), KIP‑827 (disk space metrics per log directory), KIP‑851 (requireStable flag for ListConsumerGroupOffsets), KIP‑843 (addMetricIfAbsent API), and KIP‑824 (dump‑logs segment limit).
For Kafka Streams, notable KIPs are: KIP‑846 (source/receiver node throughput metrics), KIP‑834 (pause/resume topology), KIP‑820 (merge transform() and process() APIs), and KIP‑812 (new close() API that forces members out of consumer groups).
In Kafka Connect, KIP‑618 adds exactly‑once semantics for source connectors, ensuring atomic writes of source records and offsets and preventing zombie tasks.
Beyond the listed KIPs, Kafka 3.3 includes numerous bug fixes and enhancements. For a complete list of changes, see the 3.3.0 and 3.3.1 release notes, watch the announcement video, listen to the podcast, or download Apache Kafka 3.3 and start using it.
Big Data Technology Architecture
Exploring Open Source Big Data and AI Technologies
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.