It is not so much that there are bugs in the ActiveMQConnectionFactory but that jencks is more feature rich and better able to cope with the throughput and requirement of large systems.
I don't see how the default ActiveMQConnectionFactory would not be robust enough for enterprise deployments.
However when deploying into containers you would most likely want to use some sort of connection pooling. Jencks AMQPool is one option, another is to use the ActiveMQ PooledConnectionFactory or the depending on the container the JCA compliant connection factory of ActiveMQ.
Yes we are using pooled ActiveMQConnection factories, and they are working with no problems.