Thanks Chia for review. My responses to each comment are below.. chia_0: Should we set the importance level to LOW, following the pattern of buffered.records.per.partition? >> Agree, I have updated kip.
chia_1: Is buffered.records.per.partition officially deprecated? I assumed it wasn't, but the KIP seems to suggest otherwise >> Nice catch. Indeed the config is not deprecated. (only doc str BUFFERED_RECORDS_PER_PARTITION_DOC is deprecated). I have updated kip, removing the deprecation mention. chia_2: Have you considered limiting records based on byte size rather than count? There’s a Jira (KAFKA-13152) that discussed this approach. >> Good point and agree bytes would bound memory more robustly. If we switch to bytes, configuring a good byte cap might be harder for users than configuring a records count. If we consider this kip as v1, we can perhaps file another kip for v2 based on the approach in jira 13152 (still in progress) which looks like a right long term solution applied on both these configs. Thanks, Murali On Tue, Apr 28, 2026 at 5:47 PM Chia-Ping Tsai <[email protected]> wrote: > hi Murali > > chia_0: Should we set the importance level to LOW, following the pattern > of buffered.records.per.partition? > > chia_1: Is buffered.records.per.partition officially deprecated? I assumed > it wasn't, but the KIP seems to suggest otherwise > > chia_2: Have you considered limiting records based on byte size rather > than count? There’s a Jira (KAFKA-13152) that discussed this approach. > > Best, > Chia-Ping > > On 2026/04/24 17:11:08 Muralidhar Basani via dev wrote: > > Hi all, > > > > I would like to start a discussion on KIP-1325, which introduces a new > > streams config to handle message fetching in the restore path. > > > > KIP : > > > https://cwiki.apache.org/confluence/display/KAFKA/KIP-1325%3A+Add+restore.buffered.records.per.partition+to+cap+Kafka+Streams+state-restore+memory > > > > Thanks, > > Murali > > >
