What parallel value should I set for Akka Kafka consumer when we scale out more Kubernetes pods?

I referred to this discussion to understand thread pool of Kafka consumer group of Akka cluster