In Stage/Prod, Inventory API runs with 3 replicas. While the scaling of replicas does not impact the performance of the Internal Inventory Consumer (we currently use a single partition, therefore only one consumer is processing), we should understand if there are any performance improvements or hits by having multiple replicas processing API requests from KIC

Does multiple replicas reduce the time for migration?
Does multiple replicas overload the single IIC consumer for any reason?
Are API errors more predominant due to multiple replicas?