We want to measure the maximum concurrent threads that can hit our web service before response times begin to be effected.
The best way to do that is to add one user at a time until the transactions per second tops out.
Now, what I would LIKE to do is be able to do is add a new user once per minute until the throughput for the last 5 minutes is within 1% of the throughput for the last minute, or is more than 1% higher (suggesting that performance is starting to degrade. Then I would like to be able to decide if I want to stop the scenario, or just hold the users constant at that point.
Obviously, this goes a little outside the norm. But it's doable, right?