The main thing in this release is a stable implementation of message redistribution.
This is the feature whereby messages are automatically distributed between nodes according to how fast/slow consumers are on different nodes.
I'd be particularly interested in any results in a "real" environment if anyone is willing to give it a go and experiment.
How to configure this is explained in the clustering guide inside the distribution.
Basically you just need to specify the MessagePullPolicy in the post office config as so:
You may also want to reduce the stats send period to a lower value.