I am executing a load test on an application with 2 different settings. a) With "Dynamic execute page resources" b) Keeping all the recorded page resources
The value of average throughput from both these tests are very different. What can be the possible reason for this as I can see that the requests for the resources are same in both the cases.
There's a difference when using the dynamic resources option compared to recorded resources when cache is involved. For the former the requests are not played anymore if they are in the cache whereas for the latter they are still sent and the server will respond back with a 304 not modified response.
Have you compared the number of hits on both cases? Is it roughly the same?
With dynamic extraction enabled, the resources are downloaded even when cache is enabled thus shooting the value of average throughput. On the other hand, with recorded resources being played the requests returns 304 not modified and lowering throughput. I did validate this with disabling cache.