What be the some good options to find out if a request being recieved by the dispatcher is getting served from the dispatcher or if it is travelling back to the publish instance to minimize the load on publish instances.
Increase the logging setting for dispatcher to 9 (in http.conf if you are using Apache). Then the dispatcher.log will tell you for every request whether the request is being served from cache or being sent back to the publish instances.
You don't want to leave that setting on for very long even in QA it adds quite a bit of overhead, but in a non-produciton environment it provides useful information.
In a production environment you can compare the access log of your web server to the access or request logs on your publish server to try and extract the same information.