This appears to be because the Connections are being shared and the sessions are unique. I have played with the CACHE_CONNECT, etc... Unfortunately that still appears to run synchronized spreading the load across consumers, opening and closing a connections for each message. Brutal.
Fixed, I need a beer! The problem was w/the originating messages. NMS has a
header request.Properties["NMSXGroupID"] = "NMSXGroupID" that somewhere in the depths of NMS get converted into a Group header. Each message was then forwarding to the consumer assigned the first message group.
Nothing to do w/ActiveMQ/Spring/Camel. see request.Properties"NMSXGroupID"