All LoadView subscription plans allow users to run a number of load tests of varying complexity per month. The number of load tests depends on the type of the subscription plan that have been opted for and a test configuration. In this article we will discuss how to calculate the approximate number of tests a user can run on a subscription plan. Generally, even users who need to run large load tests are able to run a few load tests within our basic Starter subscription plan.

See our recommendations on configuring your first load test in the How to Find the Web Application Breaking Point with Load Testing and Web Analytics to Concurrent Users Calculator & Formula articles of our Knowledge Base.

Information on pricing and resources included into LoadView subscription plans can be found under Account > Upgrade Plan > Load Testing Subscription.

How to Calculate the Number of Tests

First, regardless of the plan you are on, you need to configure a test device and test scenario. When the test settings are configured, go to the footer of the test scenario page. On the footer you can get an information about the amount of resources (Virtual User Minutes and Load Injector Minutes) that will be used by the system to run the test. Note that the estimated amount of the resources is based on your test settings and will be changed accordingly.

Next, find the information about your current resource balance by pointing on the resource details area. We add resources (VUM and LI Minutes) based on your subscription package to the resource balance of your account on the first day of every month. Note that all unused resources roll over to the next month.

Once you have the resources required to run the test and the balance details in hand, you are ready to calculate how many tests with the same parameters you can run. Since, running out of Load Injector Minutes (LI Min.) is the most common case, we recommend to use the LI Min. values for your calculations:

Number of Tests = Required LI Min. / Balance LI Min.

The Load Injector Minutes usage is calculated based on the number of Load Injector Servers (LIs) required for the test and the test duration time. In its turn, the number of LIs depends on the number of virtual users simulated per LI 0r, so called, Load Injector Payload. Based on our experience, by default, we use the average Payload value. However, it is recommended to calibrate your device to ensure ideal CPU utilization per LI (up to 70-80%) and optimal virtual user distribution between LIs.

Let’s see how it looks in real life. 

Browser-based (web page/web app) Load Test Calculations

Usually it is recommended to make sure a device scenario is configured correctly. To do this you can run small tests first and optimize the test settings if it is required. Then you can start large load testing of your website and combine it with small load tests every time a new feature is rolled out.

Let’s say we want to run our first load test with 25 concurrent virtual users (performing specific transactions over a specified amount of time) and 20 minutes duration.

The test with the test scenario like this will require 4 LIs to simulate the required number of virtual users. Since each LI starts at the beginning of the test and runs for 20 minutes, the total operational time of 4 LIs will be 4*20 min = 80 Load Injector Min.

We do not use Reference Load Injectors for the purposes of this article. However, it is recommended to use at least one Reference Load Injector per test to have reference test results and to record video of web application or web page load test execution.

On the Starter plan, we have 1800 Load Injector Minutes included in the plan. Let’s calculate how many tests we can run on the Starter plan based on the required LI minutes:

Number of Tests = 1800/80 = 22 load tests

If our load testing goal is 100 concurrent virtual users, the result number of tests that can be executed on the Starter plan is calculated as follows:

Number of Tests = 1800/280 = 6 load tests

HTTP-based Load Test Calculations

Since it requires less Load Injector resources to simulate HTTP virtual users, much more higher load can be handled in the case of HTTP-based load testing. For this example we set up a 20-minute HTTP load test with 1000 peak concurrent virtual users (concurrently hitting a target URL).

To run an HTTP-based test with the given configuration the system will require two LIs (the default Payload for HTTP load testing is 500 virtual users per LI). Thus, 2*20 min = 40 Load Injector Min. out of a resource balance will be used to run the test.

If we are on the Starter plan, we have 1800 Load Injector Minutes included in the plan. Let’s calculate how many tests we can run on the Starter plan based on the required LI minutes:

Number of Tests = 1800/40 = 45 load tests

In the same time, in the case of the Professional and Advanced plans we will be able to run 112 and 375 load tests with the same configuration correspondingly.