seankumar-tl commented on pull request #9319:
URL: https://github.com/apache/kafka/pull/9319#issuecomment-761571154
@rhauch @kkonstantine - This is a major piece of functionality that is
broken when running Kafka Connect at scale. Can we please get this reviewed and
merged?
-
seankumar-tl commented on pull request #9319:
URL: https://github.com/apache/kafka/pull/9319#issuecomment-731200868
@BDeus, I'm running @ramesh-muthusamy fix in production as a hotfix and can
confirm this is now the behavior. Our cluster is processing more than a million
messages a second
seankumar-tl commented on pull request #9319:
URL: https://github.com/apache/kafka/pull/9319#issuecomment-721276005
I can confirm seeing the behavior that @kkonstantine has reported. This is
with Spot instances and running multiple connectors for topics with a large
number of partitions. T