Are all the consumers in the same group? If so there could be longer
rebalance latencies; but in general you should be able to have that
Also, to which servers are you seeing the timeouts - Zookeeper or
kafka? Also, any reason why you are lowering the zk session timeout in
your code? Can you share some of your logs?
On Thu, Jan 9, 2014 at 5:07 AM, Klaus Schaefers
<[EMAIL PROTECTED]> wrote: