I don't think it really matters here because internally the source of data for us is Flux.fromIterable(consumerRecords)
which cannot lose its order because of an iterator therefore how hard we wouldn't try to process them in parallel, we still would get the order in one iterator. Yes, the order in between iterators we flatten is really unpredictable, but this doesn't matter for us since we worry about an order withing a single partition, nothing more.
Nevertheless I think we definitely need to fix that for the mentioned concatMap()
to avoid such a confusion in the future. Feel free to provide a contribution on the matter!
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…