I thought the throughput would be much better for async operations but for the given configurations, there is barely any difference between sync (30932 requests) and async (31274 requests) producers. Testing with 99 virtual users sending concurrent requests for 10 seconds to both endpoints. Topic has 3 partitions on a single broker.

I purposely force the consumer to fail consuming the message (5th) in the handler. When I restart the consumer, I don’t see any attempt to consume failed message although I didn’t call session.MarkMessage(msg, "") at the end for that message perviously. Technically the message is lost now (not visible to the consumer).