[LoadUIWeb] What is a Single User Test and why do I want to do it?
A single user test is just what it says - it is a test that uses a load of exactly one user. One option for a single user test is just to run a selected user scenario once. In fact, for convenience, WebLoadUI automatically creates a single user test for each scenario that is recorded, so that you really have no excuse not to run it!
This once-through single user test will give you a host of information that you can use to base your further analysis on. First of all, this is a sanity check - make sure your scenario runs! There may be multiple aspects to tweak, such as:
- check status of response codes. If codes are not as they were recorded you'll get a warning which you then will need to analyze and (hopefully, quickly) fix. There is a one-click option in WebLoadUI to fix warnings related to return code mismatch
- correlate your data. Even though WebLoadUI does extensive - and fast - automatic data correlation, there are unavoidably cases when additional value passing from responses to requests will need to be inserted
- look at how your pages load and what the page load times are. These are the fastest loads that you are going to get - it will go downhill as the user counts increase. The Top 10 Page reports and particularly waterfall charts are a great tool for that kind of analysis.
- Quality of Service checks. Once you've established the pages load as fast as you expect, you can attach QoS checks to flag excessive page load times with increased load.
Then, you can also run a single user test for a longer period of time, using the Continuous load option. This will help you fish out any undesirable effects related to repeatedly visiting the same pages. You think why would anything go wrong, but as the maxim goes, what you don't test does not work, so you'd better find out early than be surprised later.
After you've scrubbed your single user tests to the point of squeaky clean, you can crank up the load and compare the data that you get under load to your previous single user tests. This would be dead easy since all the reports are stored in WebLoadUI and you can look at them side by side.
This once-through single user test will give you a host of information that you can use to base your further analysis on. First of all, this is a sanity check - make sure your scenario runs! There may be multiple aspects to tweak, such as:
- check status of response codes. If codes are not as they were recorded you'll get a warning which you then will need to analyze and (hopefully, quickly) fix. There is a one-click option in WebLoadUI to fix warnings related to return code mismatch
- correlate your data. Even though WebLoadUI does extensive - and fast - automatic data correlation, there are unavoidably cases when additional value passing from responses to requests will need to be inserted
- look at how your pages load and what the page load times are. These are the fastest loads that you are going to get - it will go downhill as the user counts increase. The Top 10 Page reports and particularly waterfall charts are a great tool for that kind of analysis.
- Quality of Service checks. Once you've established the pages load as fast as you expect, you can attach QoS checks to flag excessive page load times with increased load.
Then, you can also run a single user test for a longer period of time, using the Continuous load option. This will help you fish out any undesirable effects related to repeatedly visiting the same pages. You think why would anything go wrong, but as the maxim goes, what you don't test does not work, so you'd better find out early than be surprised later.
After you've scrubbed your single user tests to the point of squeaky clean, you can crank up the load and compare the data that you get under load to your previous single user tests. This would be dead easy since all the reports are stored in WebLoadUI and you can look at them side by side.
