Is it possible to configure Apache Tomcat to simulate network latency that a client would normally have when requesting data from server for requests over localhost? I'm trying to test how the front end of a Java Servlet application will respond to requests that have slow response times.
I know I could go in and add Thread.sleep(100)
to all of my Servlet Java source files, but I'd rather have a solution that can be configured in Tomcat rather than in my actual application.
You can implement a servlet filter and have it intercept all requests. In each of the request you would be doing this Thread.sleep or something else. All requests to your servlets will be delayed.
You can simulate network latency by using a proxy on your localhost which introduces latency, bandwidth restrictions, and even drop packets.
DonsProxy will do the job. Here's a good article describing how to configure it to simulate a user on a poor network.
This is easily feasible in Tomcat by creating a valve.
Create a class that extends the ValveBase class from tomcat.
The code inside should be something like that:
Compile it in a jar you'll drop in the tomcat lib directory (usually {catalina.home}/lib).
Finally add the valve declaration in you server.xml:
HIH