I'm building a middleware component accepting requests from 2-4 instances of a frontend through a load balancer then relaying the requests to a backend, then returning the response back to the frontend servers. I've load tested with Jmeter and have noticed that the first requests from each thread takes substantially longer than subsequent requests. How can I set the keepalive timeout in Tomcat to a greater value? The keepalive timeout only works from requests from the same ip address. If this is correct, how can I utilized the keepalive timeout more effectively for requests coming from the multiple instances of the frontend? Would sticky sessions be the way to go? Thx Tom
--------------------------------- Do you Yahoo!? Yahoo! Mail SpamGuard - Read only the mail you want.
