Hi All,

I've been recently profiling the java client and noticed that it produces
considerable pressure on GC.

My main concern is the consumer as it creates multiple collections during a
single poll command: in my test system i have a consumer that polls a topic
with 100 partitions and even though no messages are coming through, the
code allocates around 100M per 5 minutes.

I've investigated the allocations and the biggest ones can be easily
avoided by moving them to the instance level, something that can be done as
KafkaConsumer is not thread safe. I've created a PR, available here
https://github.com/apache/kafka/pull/11353, that fixes the biggest
allocation.

Do you agree with the approach and if so could possibly look at my PR? Many
thanks in advance.

Regards,
Michail

Reply via email to