I have the following setup:
kafka streams application: 2 instances running
input topic: 3 partitions
output topic: 3 partitions
When I enrich the headers in the topology, I can see that partition 0 on the output topic always have record headers empty.
When I added another kafka streams instance so it matches the number of partitions on the topic, all records had headers filled in.
Anyone knows why this happens? I’m enriching through the ProcessorContext which is reached via transformValues