HiveMQ Enterprise MQTT Broker

Optimized client connection handling to reduce memory consumption.

Fixed an issue that could cause the loss of subscription data after a topology change.

Fixed an issue where enqueued publishes sent with the retain flag were mistakenly classified as retained publishes from a topic during a stateful start. This fix ensures accurate accounting and prevents possible exceptions in subsequent message flows.

Expanded the client queue state data provided in the diagnostic archive to improve debugging capabilities.

IMPORTANT: Java 21 will soon be required to run the HiveMQ Platform.

Since the HiveMQ 4.28 release in April 2024, Java 21 is recommended to run the HiveMQ Platform. For all HiveMQ versions released after April 2025, Java 21 will be required.

If you use the official HiveMQ container images, no action is required because these images have shipped with Java 21 since HiveMQ 4.28. If you do not run HiveMQ as a container, or you build your own container image, we recommend updating to Java 21 before the April 2025 deadline.

Have a great day, The HiveMQ Team

HiveMQ Team

Team HiveMQ shares deep expertise in MQTT, Industrial AI, IoT data streaming, Unified Namespace (UNS), and Industrial IoT protocols. Our blogs explore real-world challenges, practical deployment guidance, and best practices for building modern, reliable, and a secure data backbone on the HiveMQ platform, along with thought leadership shaping the future of the connected world.

We’re on a mission to build the Industrial AI Platform that transforms industrial data into real-time intelligence, actionable insights, and measurable business outcomes.

Our experts are here to support your journey. Have questions? We’re happy to help. Contact us.