Need some sort of shared cache or inter-maxscale communication for causal reads when multiple maxscales are used

Skysql currently allows for multiple maxscales to be deployed in a cluster. In the current configuration, the load balancer essentially round-robins between the two instances. As such, we believe that causal reads may not work as the queries could be split amongst the two instances.
Not sure if a shared cache, or some sort of communication between the maxscale instances would be required.