+1, this seems to be the best non intrusive option, and allows us to process a known amount of message on each poll. We can handle consumer memory footprints in a separate KIP.
-Thanks aarti On Wed, Jan 13, 2016 at 3:18 PM, Jason Gustafson <ja...@confluent.io> wrote: > Hi All, > > I'd like to open up the vote on KIP-41. This KIP adds a new consumer > configuration option "max.poll.records" which sets an upper bound on the > number of records returned in a call to poll(). This gives users a way to > limit message processing time to avoid unexpected rebalancing. This change > is backwards compatible with the default implementing the current behavior. > > Here's a link to the KIP wiki: > > https://cwiki.apache.org/confluence/display/KAFKA/KIP-41%3A+KafkaConsumer+Max+Records > > Thanks, > Jason >
