Using Ramp Sequence generator for a Peak Load Test

I think you are referring to investigating how many users there are when the average responsetime is no longer acceptable.
In that case then i recommend you to create assertions on the runner you are using and select average responsetime Then input your performance criteria, i.e the limit of responsetime which is defined as a failure.

Create a chart by selecting your runner-component in the toolbox, make sure you select average response-time and number of sent users.
Then drag-n-drop the assertions you have just created to the very same chart.

Now you should be able to investigate how many users are being sent to your service when the system is no longer responding as wanted.

These videos show you what i mean:
On Assertions
On Statistics

Let me know if you encounter any problems.

Best regards,
Mikael