The HiveMQ team is pleased to announce the availability of HiveMQ 4.4.3.

This maintenance release for the 4.4 series includes the following improvements:

• Fine-tuned timing of the HiveMQ Kubernetes Operator readiness probe to preclude premature probe completion
• Adjusted node initialization timing in the HiveMQ Kubernetes Operator to ensure steady resource utilization during rolling upgrades
• Corrected an issue that could infrequently prevent new nodes from joining a cluster
• Improved destination checks in an upstream library to ensure rolling upgrade stability
• Improved heartbeat configuration for faster failure detection
• Improved cluster merging and automatic healing capabilities for better cluster stability
• Numerous performance enhancements for cluster operations such as replication
• Reduced the amount of CPU usage and traffic that client disconnection generates when a node in a HiveMQ cluster stops
• Corrected an issue that could disrupt message flow under some circumstances
• Fine-tuned the dns-4.4.3 Docker image for accurate configuration of the cluster transport type
• Fixed an issue where the Prometheus extension did not return metrics under load
• Corrected format error in the heap-dump path to ensure that a heap dump is automatically generated when HiveMQ runs out of memory
• Adjusted the way that license files are read to the license folder to increase verification accuracy
• Fixed a rare NullPointerException that could appear during the license check of a cluster

If you currently use HiveMQ 4.x, we recommend upgrading to version 4.4.3.

Have a great day, The HiveMQ Team