In the initial "How to Move IBM App Connect Enterprise to Containers" post, a single MQ queue was used in place of an actual MQ-based back-end system used by an HTTP Input/Reply flow, which allowed for a clean example of moving from local MQ connections to remote client connections.
In this post, we will look at what happens when an actual back-end is used with multiple client containers and explore solutions to the key challenge: how do we ensure that reply messages return to the correct server when all the containers have identical starting configurations?