Hi folks, We have made a small change to this KIP for the Apache Kafka 3.2 release.
When we tried to use this in Kafka Streams, we noted a significant degradation of the performances, see [1]. It is not clear whether the prefixing is the root cause of the issue or not. To be on the safe side, we have removed the prefixing. It does not bring much anyway as we are still able to distinguish a custom reason from the default one. [1] https://github.com/apache/kafka/pull/11873 Best, David On Wed, Dec 1, 2021 at 2:17 PM David Jacot <[email protected]> wrote: > > Hi all, > > With 4 binding +1 votes (Gwen, Mickael, Tom, Bill) and 1 > non-binding +1 vote (Luke), the vote passes. > > Thanks everyone! > > Best, > David > > On Fri, Nov 26, 2021 at 4:43 PM Bill Bejeck <[email protected]> wrote: > > > > Thanks for the KIP, David this seems like it will be very helpful. > > > > +1(binding) > > > > -Bill > > > > On Thu, Nov 25, 2021 at 10:02 AM David Jacot <[email protected]> > > wrote: > > > > > Hi Tom, > > > > > > I do agree with you. For context, this is the current reason/message > > > logged > > > by the consumer when enforceRebalance is called so I just kept it as it > > > is. > > > I guess that we can revise it during the implementation. > > > > > > Thanks, > > > David > > > > > > On Thu, Nov 25, 2021 at 3:52 PM Tom Bentley <[email protected]> wrote: > > > > > > > > Thanks for the KIP David. > > > > > > > > It's a very trivial point, but I did wonder whether "caused" or > > > "requested" > > > > might be a better word than "enforced" in the default messages. > > > > > > > > In any case, +1 (binding). > > > > > > > > Kind regards, > > > > > > > > Tom > > > > > > > > > > > > > > > > On Thu, Nov 25, 2021 at 2:36 PM David Jacot <[email protected] > > > > > > > > wrote: > > > > > > > > > Thanks, Mickael. I have removed it. > > > > > > > > > > On Thu, Nov 25, 2021 at 3:22 PM Mickael Maison < > > > [email protected]> > > > > > wrote: > > > > > > > > > > > > +1 (binding) > > > > > > Thanks for the KIP > > > > > > > > > > > > Just a small suggestion: in other Options classes we don't seem to > > > > > > prefix setters with "set", so I think we could use void reason(final > > > > > > String reason). > > > > > > > > > > > > > > > > > > On Thu, Nov 25, 2021 at 1:48 PM Luke Chen <[email protected]> wrote: > > > > > > > > > > > > > > Hi David, > > > > > > > Thanks for the KIP! > > > > > > > It is good to have the joinGroup/leaveGroup reason sent to brokers > > > for > > > > > > > better troubleshooting. > > > > > > > > > > > > > > +1 (non-binding) > > > > > > > > > > > > > > Thank you. > > > > > > > Luke > > > > > > > > > > > > > > On Thu, Nov 25, 2021 at 8:14 AM Gwen Shapira > > > <[email protected] > > > > > > > > > > > > > wrote: > > > > > > > > > > > > > > > +1 > > > > > > > > > > > > > > > > Thanks for driving David. Super useful. > > > > > > > > > > > > > > > > On Wed, Nov 24, 2021 at 8:53 AM David Jacot > > > > > <[email protected]> > > > > > > > > wrote: > > > > > > > > > > > > > > > > > Hi folks, > > > > > > > > > > > > > > > > > > I'd like to start a vote on KIP-800: Add reason to > > > > > LeaveGroupRequest. > > > > > > > > > > > > > > > > > > KIP: https://cwiki.apache.org/confluence/x/eYyqCw > > > > > > > > > > > > > > > > > > Please let me know what you think. > > > > > > > > > > > > > > > > > > Best, > > > > > > > > > David > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > > > > Gwen Shapira > > > > > > > > Engineering Manager | Confluent > > > > > > > > 650.450.2760 | @gwenshap > > > > > > > > Follow us: Twitter | blog > > > > > > > > > > > > > > > > > > > > >
