Ask HN: Benchmarking Kafka Request Reply

I have inherited an application which uses Kafkas request/reply configuration to provide a (mostly) synchronous configuration, which seems to work well enough. However I have now been approached by management to figure out what sort of scale this can handle. I am a bit green with Kafka so did a bit of looking around, however my normal load testing tools seem to not support Kafka request reply.

Gatling doesn't want to cover this, and points to a third party plug in, which appears to be undocumented and archived.

JMeter can generate load via Pepperbox, but doesn't consume.

There appears to be a lot of information on generating load, but nothing about measuring throughput in this set up. Is this a step too far in validating an asynchronous service, or are there setups that can be used to understand the limits of our infrastructure?