My application calls 10-20 web services in parallel to serve a particular user functionality using multiple threads. These web services usually respond within 5-45 seconds depending on the nature of the call and service. However my server alarms are showing a high tcp re-transmission percentage. But the time outs or non results from these web services are very rare.
My questions are:
1. Is these delays cause the high % as the server assumes packets are lost ?
2. Does these retransmission rate affect the performance of my server ?
3. Is there any workaround for the same ?