I have a virtual service that uses a Live Invocation step as part of the model. I'm building it as a proof of concept to support routing traffic between various environments but it also needs to hold up under load.
My model looks like this (very simple) -
Listener -> Live Invocation -> Responder
If I use SOAPUI to hit the Live Invocaation URL directly (from the server that is hosting DevTest), I get a <100ms response time but if I go through DevTest, my response time jumps to ~1 sec (1000 ms), almost 10x.
Are there any methods to optimize this step? Is there a concept of connection pooling or reuse?