throghput vs. load

I assume this means that my scenario would have 10 users per second when we hit 20 minutes (or does it involve that my SUT responses takes longer time)?
I use a .forverer around this load model.

`

(rampUsersPerSec(1) to (10) during (20 minutes))

`

That will 10 users in 20 minutes. 1 user about every 2 minutes.

All your simulation setup questions are explained here too.
http://gatling.io/docs/2.0.3/general/simulation_setup.html

At the start of the test it will inject 1 user per second. This injection rate will increase linearly to 10 at the end of the test.

I am not convinced that you need your users to loop forever -any reason for that? Thanks

Thanks, no. No reason. I will remove it.