"In order to process messages that are received concurrently, you can configure additional instances on the KafkaConsumer node. When additional instances are configured, a single Kafka message consumer is created, and the messages are distributed to the additional flow instances. As messages are processed concurrently, message ordering is not preserved when additional instances are being used."
— IBM Docs - https://www.ibm.com/docs/en/app-connect/13.0.x?topic=enterprise-consuming-messages-from-kafka-topics
What this means:
ACE creates 1 Kafka consumer only, regardless of the number of additional instances. This one Kafka consumer receives all messages from the topic (single partition in your case).
ACE then internally distributes those messages across the available flow instances (including the additional instances). This leads to parallel processing, despite the Kafka topic having only one partition.
However, this breaks message ordering because messages are processed out-of-order by different instances.