Saltar al contenido

Black Friday: How to plan your stress tests with JMeter

A new Black Friday is coming November 27th of this year. Are you ready?

One of the main activity to assess is the responsiveness of the ecommerce application under stress. This means, keeping the proper level of operation and performance of your ecommerce services under unusually high traffic.

The way to empirically assess this very important point is the emulation of the high peaks of simultaneous users accessing the ecommerce in a parallel environment (usually pre-production or QA).

What are the key points of stress testing?

During the simulation of high traffic, it is crucial to monitor the systems and the applications in order to understand:

  1. How many simultaneous users can the platform/infrastructure support while maintaining optimal performance (load test)
  2. How many simultaneous users can the platform/infrastructure support before the system starts to show signs of failure (stress test).

How can it be performed?

Load and stress testing should be implemented carefully.

The most adopted framework is Apache JMeter. However, a simple instance of JMeter cannot be enough if you need to test more than 500 simultaneous users. In these cases a more complex configuration is needed or eventually a different approach.

Are there any SaaS applications that can perform stress testing?

One of the best Software-as-a-service (SaaS) around, that is able to take JMeter test files and ejecute as many simultaneous users as you need is BlazeMeter. However, you should be ready to face the cost: it is not so cheap.

What are the alternatives?

In Orienteed, we have developed a JMeter cluster solution using AWS (Amazon Web Service). These tests are developed and run within our infrastructure.

We are able to execute tests with more the 10 thousands simultaneous users providing readable reports focusing on the most critical services affecting the performance.

Below is a table showing an example of ecommerce page transactions ordered by Average Load Time:

Load time per transaction

In addition, with Orienteed's ELK instance (Elastic, LogStash, Kibana) we can easily monitor any ecommerce ecosystem and application.

We have already developed an APM solution for HCL Commerce, Hybris Commerce and Magento.

ELK APM Dashboard

If you would like to learn more about our quality assurance testing and monitoring solution, we highly recommend that you contact our team of specialists to evaluate your ecommerce.

At Orienteed, we can ensure that your ecommerce is operating optimally at 100% capacity during periods of strong online sales such as Black Friday.