Justine, Thanks for bringing this up, I checked the two KIPs and I think KIP-679 subsumes KIP-185 so I can discard the latter.
As for the configs of "max.in.flight.requests.per.connection" and "retries": I think it's okay to keep the default as 5 for now since on the broker side we still keep a history of sequences, as for "retries", I think that's already replaced by the newly introduced "max.block.ms". Guozhang On Wed, Jul 13, 2022 at 8:56 AM Justine Olshan <jols...@confluent.io.invalid> wrote: > Hey Apache Kafka community > > I was doing some research on idempotent producers and found KIP-185 > < > https://cwiki.apache.org/confluence/display/KAFKA/KIP-185%3A+Make+exactly+once+in+order+delivery+per+partition+the+default+producer+setting > > > . > > It seems like this KIP was replaced by KIP-679 > < > https://cwiki.apache.org/confluence/display/KAFKA/KIP-679%3A+Producer+will+enable+the+strongest+delivery+guarantee+by+default > >. > There are minor differences between the KIPs, but I'd say KIP-679 replaces > KIP-185. > > Do we think it is valid to discard KIP-185? Or at least move it from "under > discussion" on the KIP page > < > https://cwiki.apache.org/confluence/display/KAFKA/Kafka+Improvement+Proposals > > > ? > > Thanks, > Justine > -- -- Guozhang