I tested with 1 connection with 10 threads and that was running fine. 1 connection with 15 or more threads seems to be an issue with reading from TCP and servicing all the threads??
In our test we tried with 6 connections and we were able to just use 25 threads. For our app we needs to service 100's of threads. For now we are use semaphore to limit number of threads calling send api but the latency is an issue as many threads are trying to wait for few connections (10-25 depending on # of connections). -- View this message in context: http://activemq.2283324.n4.nabble.com/FutureResponse-bottleneck-when-using-more-than-10-threads-tp4702945p4703119.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
