>There is also a general expectation that there is strict order - events should be read in the same order that they arrived in the system.
This is the Achilles' heel of Kafka (and Pulsar) where for all streaming systems (or workarounds) with per message key acknowledgements incur O(n^2) costs in either computation, bandwidth, or storage per n messages [1].
[1] What If We Could Rebuild Kafka from Scratch?: (Top comment of the 220 comments)
This is the Achilles' heel of Kafka (and Pulsar) where for all streaming systems (or workarounds) with per message key acknowledgements incur O(n^2) costs in either computation, bandwidth, or storage per n messages [1].
[1] What If We Could Rebuild Kafka from Scratch?: (Top comment of the 220 comments)
https://news.ycombinator.com/item?id=43790420