Hi folks, Friendly ping on this KIP.
Thanks, Aditya > On May 11, 2026, at 18:12, Aditya Kousik <[email protected]> wrote: > Hi all, > One more nudge on KIP-1307. > > Interceptor exceptions are logged today, but in a high-throughput client, > intermittent interceptor failures drown in noise/subject to log rotation; and > a slow interceptor is invisible to the client. A metric is the right > primitive for this scenario. > > Looking for feedback on the viability of this option. > > Thanks and regards, > Aditya > >> On Apr 16, 2026, at 12:18, Aditya Kousik <[email protected]> wrote: >> >> Hi everyone, >> >> Quick follow-up on this KIP. Mickael included a nice summary in the March >> 2026 Monthly Digest: >> https://developers.redhat.com/blog/2026/04/03/kafka-monthly-digest-march-2026#kafka_improvement_proposals >> >> Appreciate any feedback and thoughts you might have. >> >> Best, >> Aditya >> >>>> On Apr 1, 2026, at 08:34, Aditya Kousik <[email protected]> wrote: >>> Hi all, >>> I'd like to start a discussion on KIP-1307: SerDe and Interceptor Metrics. >>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-1307%3A+SerDe+and+Interceptor+Metrics >>> Serialization, deserialization, and interceptors run in the client hot path >>> with no timing or error metrics. A slow custom SerDe or interceptor is >>> currently indistinguishable from a slow broker. This KIP adds the missing >>> metrics using the existing Sensor infrastructure. >>> Looking forward to your feedback. >>> Thanks, >>> Aditya Kousik
