we have a single partition, and the topic contains 300k events. we fired off a camus job, it finished within 1 minute. this is rather fast.
I was guess that the multiple mappers must be reading from multiple offsets in parallel, right? otherwise if they are reading in serial (like in a consumer group, which a previous thead has clarified to be not the case, i.e. camus does not use group api), it would be very slow. Thanks Yang
