To ensure High Availability in a multiple ProcessRobot Servers configuration, the use of Redis Servers is necessary.
Suppose a Load Balancer only configuration between ProcessRobot Servers is made (without Redis).
In this configuration, High Availability will not be possible (as a Client-Server system setup is only as available as its weakest link) - which in this case is a problem since the communication between multiple PR Servers is not handled when not using Redis Server(s). For example, if a PR Server fails all information will be lost as it will not be stored in Redis.
In other words, the use of a Load Balancer is only responsible for achieving a scalable infrastructure (meaning that communication between client tools and PR Servers will be managed when server capacity and network limits are reached/exceeded).
However, in order to create a High Availability setup, you would need Redis servers (in a master-slave mode through replication) to ensure the integrity of ProcessRobot servers state, and thus ensuring the uninterrupted execution of processes in case of a disaster (as per below image).