Hi Ryan, Thanks for your feedback. I’ll reply in the discussion thread.
Best Regards, Jiunn-Yang > Ryan Leslie (BLOOMBERG/ NEW YORK) via dev <[email protected]> 於 2026年5月15日 > 凌晨12:55 寫道： > > Hi Jiunn-Yang, > > Thanks for the new KIP. I like that the approach addresses KIP-1282 concerns > and is simplified to focus on hot partitions only. > > One question is if the config property would be better named as > auto.offset.reset.latest.min.age.ms instead of auto.offset.reset.max.age > > .ms makes the option consistent with others such as default.api.timeout.ms > and connections.max.idle.ms, and the 'ms' is good in general to communicate > the expected unit. > > The reason I am thinking 'min' instead of 'max' is the latest option will > apply only if the partition age is at least this value (at a minimum), > otherwise it will be earliest. So it is min age for latest, max age for > earliest, but the config is attached to latest. > > -Ryan > > From: [email protected] At: 05/14/26 07:47:27 UTC-4:00To: > [email protected] > Subject: [VOTE] KIP-1327 Prevent Hot Data Loss on Partition Expansion for > Latest Policy > > Hello all > > I would like to call a vote on KIP-1327 Prevent Hot Data Loss on Partition > Expansion for Latest Policy > > KIP: > <https://cwiki.apache.org/confluence/x/KY4mGQ> > > Discussion thread: > <https://lists.apache.org/thread/bp4zk31zr1sdxjsspg7b7bqddmm9t4gn> > > Best Regards, > Jiunn-Yang >
