Hi Yaroslav, +1 from me. It is meaningful to keep the common Kafka client configuration the same between Kafka source and sink.
Best, Hang Mason Chen <mas.chen6...@gmail.com> 于2022年10月13日周四 13:51写道： > Hi Yaroslav, > > +1 from my end too. I get some questions internally on the warning logs due > to conflicting client-id and it would be nice to resolve them. > > Best, > Mason > > On Wed, Oct 12, 2022 at 2:03 PM Martijn Visser <martijnvis...@apache.org> > wrote: > > > Hi Yaroslav, > > > > +1 from my end. Thanks for bringing this up! > > > > Best regards, > > > > Martijn > > > > On Wed, Oct 12, 2022 at 6:34 PM Yaroslav Tkachenko > > <yaros...@goldsky.com.invalid> wrote: > > > > > Hi everyone, > > > > > > I'd like to propose adding client.id.prefix to the KafkaSink to mirror > > the > > > functionality provided by the KafkaSource. > > > > > > Defining client.id is very important when running workloads with many > > > different Kafka clients: they help with identification and enforcing > > > quotas. Due to the specific implementation details, you can't simply > set > > > client.id for a KafkaSource or KafkaSink in Flink, you can find more > > > context in https://issues.apache.org/jira/browse/FLINK-8093, which was > > > opened in 2017 and is still unresolved. > > > > > > I don't see any reason to treat KafkaSink differently: if we can define > > > a client.id.prefix for the KafkaSource we should be able to define it > for > > > the KafkaSink. > > > > > > I've created an issue > https://issues.apache.org/jira/browse/FLINK-28842 > > > and > > > a PR https://github.com/apache/flink/pull/20475 to resolve this. > > > > > >
