Instaclustr for Apache Kafka® and Apache Kafka® Connect 3.9.2 are Generally Available

April 1, 2026 | By Varun Ghai

 

We are pleased to announce that Apache Kafka® 3.9.2 and Apache Kafka® Connect 3.9.2 are now generally available (GA) on the NetApp Instaclustr Managed Platform. 

Kafka 3.9.2 is a patch (bug‑fix) release for the 3.9.x line, delivering important stability improvements, security updates, and other fixes across Kafka brokers, controllers, clients, and Connect. In total, Kafka 3.9.2 fixes 35 issues compared to Kafka 3.9.1, making it a recommended upgrade for customers running earlier 3.9.x versions. Full details of the upstream changes are available in the official release notes and release announcement.  

As with all GA releases on our platform, this version has been validated to meet our operational, reliability, and support standards before being made available to customers. 

 

What’s New in Apache Kafka 3.9.2 

Aside from security upgrades and upgrades of third‑party dependencies to address reported CVEs, including updates to commonly used libraries, some of the more noteworthy inclusions are: 

Improved consistency between ZooKeeper and KRaft modes
This release includes KIP‑1252, which addresses inconsistent behavior when using AlterConfigPolicy between legacy ZooKeeper‑based clusters and KRaft‑based clusters. Kafka 3.9.2 introduces an optional configuration that allows ZooKeeper‑mode clusters to opt into KRaft‑compatible behavior, helping operators achieve more consistent policy validation across deployment modes.  

Broker, controller, and metadata fixes
Multiple fixes improve broker stability, metadata handling, controller behavior, and log management, reducing the risk of edge‑case failures in busy or long‑running clusters. 

  • KAFKA‑19026 – AlterConfigPolicy incompatibility between ZooKeeper and KRaft modes
    Fixes inconsistent behavior when using AlterConfigPolicy with SUBTRACT and APPEND operations, and introduces the KIP‑1252 compatibility option to align ZooKeeper behavior with KRaft expectations.
    https://issues.apache.org/jira/browse/KAFKA-19026  
  • KAFKA‑19719 – –no-initial-controllers flag validation issue during KRaft startup
    Prevents incorrect validation of KRaft metadata versions when the --no-initial-controllers flag is used, improving reliability during cluster initialization and migration scenarios.
    https://issues.apache.org/jira/browse/KAFKA-19719  
  • KAFKA‑19334 – MetadataShell bypasses file lock due to lock file deletion
    Fixes a bug where MetadataShell could unintentionally bypass locking guarantees, improving metadata safety and preventing concurrent access issues.
    https://issues.apache.org/jira/browse/KAFKA-19334  
  • KAFKA‑17076 – logEndOffset could be lost during log cleaning
    Addresses a correctness issue where log cleaning could result in loss of logEndOffset, improving broker log integrity and recovery behavior.
    https://issues.apache.org/jira/browse/KAFKA-17076 

Client and protocol correctness
The release includes fixes for producer, consumer, and group coordination edge cases, improving correctness during rebalances, offset fetches, and transactional workflows.  

  • KAFKA‑19439 – OffsetFetch API does not return group‑level errors correctly
    Corrects OffsetFetch API behavior so group‑level errors are returned properly, improving protocol correctness for consumers and monitoring tools.
    https://issues.apache.org/jira/browse/KAFKA-19439 
  • KAFKA‑19479 – Kafka Streams at‑least‑once mode could silently drop messages
    Fixes a violation of delivery guarantees where messages could be dropped if the producer failed with MESSAGE_TOO_LARGE.
    https://issues.apache.org/jira/browse/KAFKA-19479 
  • KAFKA‑19242 – Commit bugs caused by race conditions during rebalancing
    Fixes race conditions that could cause incorrect offset commits during consumer group rebalances, improving delivery guarantees and group stability.
    https://issues.apache.org/jira/browse/KAFKA-19242 
  • KAFKA‑19561 – Request timeout during SASL reauthentication
    Addresses request timeouts caused by missed OP_WRITE interest registration during SASL reauthentication, improving client connectivity stability.
    https://issues.apache.org/jira/browse/KAFKA-19561 

 

Lifecycle Guidance and Upgrade Recommendations 

With the release of Kafka 3.9.2, we have reviewed and updated the lifecycle states of older Kafka versions in accordance with our published lifecycle policy. To ensure continued access to full support and SLAs, we recommend that customers: 

  • Run GA versions of Kafka and Kafka Connect wherever possible 
  • Upgrade Kafka and Kafka Connect clusters at least once per year to a current GA release 
  • Plan upgrades proactively, especially when running older patch versions 

Before upgrading production clusters, we strongly recommend validating Kafka 3.9.2 in a non‑production environment to confirm compatibility with your Kafka clients, connectors, and applications. 

Details of all Kafka versions currently supported on the Instaclustr Managed Platform are available in our documentation. 

 

Getting Started or Upgrading 

If you need any help spinning up a new Kafka cluster or upgrading existing managed Kafka clusters, please feel free to get in touch with us via our support website.