1、Start Vuser
ep1： Strat 100 Vusers :2 every 00:00:15(HH:MM:SS)
Explanation: A total of 100 virtual users need to be run in the scenario, and 2 virtual users Vusers are started every 15 seconds. It takes 12 minutes and 30 seconds to start 100 virtual users in total.
ep2：Strat 100 Vusers simultaneously
Explanation: The scene dispatches 100 virtual users at the same time when it starts running.
2. Duration
ep: Run for 00:05:00（HH:MM:SS）
The virtual user in the scene needs to continuously execute for a total of 5 minutes in the scene. (If the recording time only takes 10 seconds to execute once, then one Vuser executes dozens of times within 5 minutes at a certain time, and the number of events for 100 Vusers reaches several thousand within 5 minutes);
3、Stop Vuser
ep1：Stop all Vusers : 5 every 00:00:30(HH:MM:SS)
Explanation: When the Vuser runs out, how to stop all the Vusers in the scene? Here, it is set to stop 5 Vusers every 30 seconds, and 100 Vusers need 10 minutes to stop all.
ep2：Stop all Vusers simultaneously
Explanation: After running the scenario, set to stop all running Vusers at the same time.
The Interactive Schedule Graph on the right side of the scene settings reflects the settings on the left side, which can be understood by comparison.
In the
performance
, the most critical step is the design of the Controller scene, because the design of the scene
is
related to the design of the test case, and the execution of the test case directly affects the final test result. Designing a scenario may be the execution of a test case (there can be multiple scripts in a scenario design, and the scenario planning method can be in groups or scenarios). If the scenario design is incorrect or unreasonable, Then there is no point in analyzing the results in Analysis, right?
Let's share the issues that need to be paid attention to and understood when designing the scene in the Controller:
1. The duration setting in the scene will override the Vuser iteration setting. This means that if you set the duration to 5 minutes, then the Vuser will continue to run as many iterations as possible within the 5 minute period, even if the runtime setting only specifies 1 or 2 iterations.
2. The number of initialization Vuser activities in the scenario global plan affects the timeout value. For example, 100 Vusers trying to initialize will take longer than 10 Vusers trying to initialize.
LoadRunner
will add an internal value to the specified timeout value based on the number of active Vusers.
3. VuGen will not execute the lr_think_Times function during playback in the script, because this will cause more pressure on the server. It is recommended to set a reasonable thinking time in the Run-time settings, generally 3 to 5 seconds.
4. Whether or not to add a rendezvous point and the rendezvous point strategy in the scene will affect the performance test results more or less (the premise is that there is a function to add a rendezvous point in the script). The data such as number of clicks", "total number of clicks", "throughput" will be more than when no rendezvous point is added, and the response time is relatively realistic and can
reflect
the effect of stress test, especially when the number of users is relatively large Do concurrency.