[
https://issues.apache.org/jira/browse/KAFKA-10847?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17271698#comment-17271698
]
A. Sophie Blee-Goldman commented on KAFKA-10847:
------------------------------------------------
Hey [~spena] – to answer your question at the end there, yes, a particularly
long punctuation could definitely cause the thread to drop out of the consumer
group and trigger a rebalance. But the default (and probably minimum of any
serious application) max.poll.interval.ms is 5 minutes, so the punctuation
would have to be pretty long to cause this. If the punctuations are getting
into the range of multi-minutes and blocking all further work on that thread,
then I would say that's a problem in its own right, regardless of the
rebalancing.
By the way, given the potential for performance regressions here I think it
would be best if you could run some benchmarks on a basic POC with the above
implementation. If it does seriously impact the performance then we can
consider other designs, or possibly make this feature opt-in/opt-out which
would then require a KIP
> Avoid spurious left/outer join results in stream-stream join
> -------------------------------------------------------------
>
> Key: KAFKA-10847
> URL: https://issues.apache.org/jira/browse/KAFKA-10847
> Project: Kafka
> Issue Type: Improvement
> Components: streams
> Reporter: Matthias J. Sax
> Assignee: Sergio Peña
> Priority: Major
>
> KafkaStreams follows an eager execution model, ie, it never buffers input
> records but processes them right away. For left/outer stream-stream join,
> this implies that left/outer join result might be emitted before the window
> end (or window close) time is reached. Thus, a record what will be an
> inner-join result, might produce a eager (and spurious) left/outer join
> result.
> We should change the implementation of the join, to not emit eager left/outer
> join result, but instead delay the emission of such result after the window
> grace period passed.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
