Transactions over a period of time
When performance testing, a lot of time gets spent calibrating your tests. To do this effectively, you often have to calibrate using multiple methods. One method I use is to look at transactions over a period of time.
When looking at transactions over a period of time, your concern is that your test isn't doing too much or too little over the time interval. Examples might include logins/logouts per minute, form or file submissions rates per minute, reports generated per minute, web service calls per second, searches per second, etc. The idea is that you are looking to roughly approximate load (as determined by transactions) with your test.
Likely, you won't just be looking at one transaction. You'll be looking at many of them. It's a bit of a balancing act. Tuning one might knock one of your other transactions out of whack. You then get to tune again.
You'll need to have some idea of what these transactions might look like in production (both actual and forecasted). You might decide to build a couple of different scenarios based on the data you get. If your tool alows it, try to build in a way to programatically track and throttle these numbers as needed. It might save you a lot of time.
When looking at transactions over a period of time, your concern is that your test isn't doing too much or too little over the time interval. Examples might include logins/logouts per minute, form or file submissions rates per minute, reports generated per minute, web service calls per second, searches per second, etc. The idea is that you are looking to roughly approximate load (as determined by transactions) with your test.
Likely, you won't just be looking at one transaction. You'll be looking at many of them. It's a bit of a balancing act. Tuning one might knock one of your other transactions out of whack. You then get to tune again.
You'll need to have some idea of what these transactions might look like in production (both actual and forecasted). You might decide to build a couple of different scenarios based on the data you get. If your tool alows it, try to build in a way to programatically track and throttle these numbers as needed. It might save you a lot of time.