-1 Let me make an observation and then state my reasons.
First, I am a little surprised to see this vote thread as I think it is a bit premature. As a community, we've consistently invested in good communication. If there are differing opinions on a DISCUSS thread, we generally gather the community for a KIP call/ in-person meeting. If required, meet several times. Then send around notes and finally, call a vote. I've observed us resolve many issues and reach better decisions as a community by doing that. We've given higher bandwidth communication a chance and that has paid off. I would've liked to see us do the same here and would still offer the time to do so. There is a ton of context involved to understand the concerns of those of us who have been actively contributing to Kafka for 6 years -- We've worked very hard to keep Kafka simple, that is a big reason for its adoption and success so far. We've built a thoughtful community that engages in constructive communication versus reaching knee-jerk decisions. I see the latter happening here and would urge the new members to invest more time in sharing ideas before voting. Let me state the reasons for my vote since the thread was started: - REST proxy does not cover a meaningful % of usage on Kafka. When we ran a poll, it turned out that 50% users want non-java clients, 46% ish wanted the java clients and the rest were interested in other protocols (REST, MQTT, AMQP). - We have tried this approach and it did not work. Our objective is to deliver high-quality software and build an agile community. We failed at both those objectives with clients and have decided to instead invest in the ecosystem and not bloat Apache Kafka. - The Kafka ecosystem is large, thriving, open-source and non-Apache. That is working, why do we think this should be an exception? Never change something that works right? To me, this already is a big reason why the Apache argument doesn't make sense. Regarding the governance worries -- If you are worried about governance, why not first try to contribute, hit roadblocks around project direction and then raise concerns around governance? Doing that prematurely is solving a hypothetical problem and we need to invest time in solving real problems. On Tue, Oct 25, 2016 at 5:30 PM Haohui Mai <ricet...@gmail.com> wrote: > +1 > > In my personal experience Kafka REST server is one of essential parts on > building a reliable real-time pipeline. It's okay to keep the code > reasonably separated from the core, but would appreciate to see some > official supports for the REST servers from the Kafka community. > > Regards, > Haohui > > On Tue, Oct 25, 2016 at 4:42 PM Ismael Juma <ism...@juma.me.uk> wrote: > > > -1 > > > > I think it's good to avoid adding non-essential items to the Apache Kafka > > project so that it can continue evolving at a fast pace. Since a REST > > server can be naturally implemented as an external project, I believe the > > disadvantages of including it outweigh the advantages. > > > > Ismael > > > > On Tue, Oct 25, 2016 at 10:16 PM, Harsha Chintalapani <ka...@harsha.io> > > wrote: > > > > > Hi All, > > > We are proposing to have a REST Server as part of Apache > > Kafka > > > to provide producer/consumer/admin APIs. We Strongly believe having > > > REST server functionality with Apache Kafka will help a lot of users. > > > Here is the KIP that Mani Kumar wrote > > > https://cwiki.apache.org/confluence/display/KAFKA/KIP- > > > 80:+Kafka+Rest+Server. > > > There is a discussion thread in dev list that had differing opinions on > > > whether to include REST server in Apache Kafka or not. You can read > more > > > about that in this thread > > > > http://mail-archives.apache.org/mod_mbox/kafka-dev/201610.mbox/%3CCAMVt_ > > > aymqeudm39znsxgktpdde46sowmqhsxop-+jmbcuv7...@mail.gmail.com%3E > > > > > > This is a VOTE thread to check interest in the community for > > > adding REST Server implementation in Apache Kafka. > > > > > > Thanks, > > > Harsha > > > > > > -- Thanks, Neha
