Configuring a stress test requires you to set up a test load that reflects real-world behavior. Put another way, to receive the most relevant test results, you need to specify the realistic number of concurrent users to simulate during the test at the target web resource.
Concurrent users represent how many virtual users will visit your website and perform specific transactions at predefined regular time interval (but not the same step of the transaction all at once).
Using web analytics tools enables you to find out relevant traffic data, such as user visits statistics and duration of sessions on a website.
Using Web Analytics
Google Analytics, as well as some other web analytics tools, provides you with the Visits metric your website has per a regular time interval and the Average Visit Duration for the selected interval. To translate these metrics into concurrent users, you need to find Hourly Visits statistics and Average Visit Duration for the same time period.
Use the following formula to estimate the number of concurrent users:
Concurrent users = Hourly Visits x Average Visit Duration (in minutes)/60
While the average traffic statistics can work well to set up a realistic test, it is more significant to use statistics of peak traffic periods to configure a test.
For example, during peak hours of traffic, your site analytics shows the peak number of Visits per hour equal to 1000 visitors. Let’s say the Average Visit Duration value for the interval is 5 minutes.
In this case, the number of concurrent users that should be simulated to create 1000 visits per hour is:
Concurrent users = 1000 x 5/60 = 83
Using Custom Parameters
If you don’t have web analytics data, you can use the expected number of user visits to calculate the number for concurrent users:
Concurrent users = Number of expected visits per minute * Visit Duration (in minutes)
Let’s say we expect 100,000 users to visit our website per day. This is equal to 100,000/24 = 4,166.66 per hour. For this example, we’ll approximate to 4,200 per hour. Next, to find the number of visits per minute, we take 4,200/60 = 70 visits per minute. If the predicted visit duration is 30 seconds, or 0.5 minutes, the number of concurrent users that should be simulated to create 70 visits per minute is:
Concurrent users = 70*0.5 = 35
This means a single virtual user can run 2 session per minute (1 minute/0.5 minute = 2 sessions per minute). In this case, we need LoadView to simulate 35 virtual users to emulate 70 visits on the target website during a minute.