During the conference Ray mentioned that it would be really nice to have Openshift autoscaling policies for Infinispan based on:

• number of entries (and/or)
• memory usage

Some examples might be found in his repos