When and How We Use Load Testing

For stable peak performance, crisis prevention, and risk management, it is important to keep your IT system resilient. The load testing of the website or application will help to ensure its stable operation.

We want to share our experience and explain how we investigate ‌performance, identify and eliminate bottlenecks, and determine the fault tolerance of the system.

Load Testing Stages

Load testing is part of the system performance audit. First, we define the goals, describe the required load indicators, and identify the issues. This is a study which involves the collaboration between all team members, including the developer, the architect, and DevOps.

Next, we proceed directly to ‌load testing. As a result, we generate a report with data on peak performance, key performance indicators, time of responses to user requests, system resource utilization, and description of critical resources.

Based on this report, we provide our recommendations. By following them, our customers can themselves optimize the performance or assign this process to us.

The works are implemented in accordance with the identified issues. For example, this could involve ‌code optimization or eliminating bottlenecks. The list of specific tasks depends on the test results. We also run tests to determine the effectiveness of these changes.

Let's discuss a case where an IT system needs load testing.

Dealing with a High Volume of Users

If you are preparing a large-scale advertising campaign, planning to centralize or merge systems, expect a surge in demand, your IT product will have to deal with a high volume of customers. It is a good practice to identify in advance your system limits in terms of users. This is where load testing could be helpful.

What you get:

• identifying the capacity and limits of acceptable load;
• ensuring advance detection of problem areas.

Our projects involving ‌load testing included a major fast food restaurant. The partner forecasted an upsurge in mobile traffic. In two weeks, we ensured the stable operation of the app by doubling the system resiliency.

Saving an IT System at a Critical Point

If you skip performing ‌load testing in advance, your product may not be able to deal with a high volume of users. This can damage your business reputation and erode the loyalty of your customers. The reasons may include the lack of hardware resources, non-optimized code, or obsolete hardware.

What you get:

• optimizing system performance;
• preserving user traffic;
• mitigating reputational damage.

We addressed such issues faced by one of our customers following a promotional campaign. The upsurge in users led to a performance glitch in the mobile app. This increased the loading time on devices and, when the response time was more than 1 minute, the app would display a message that the server is unavailable. As the issue caused frustration among the users, the situation required them to take some urgent steps.

We promptly responded to the incident and, within two days, optimized API queries and restored the app's operation. This resulted in a 15-fold performance increase.

Implementing Continuous Optimization

The smooth operation will be of no help to your product if you ignore regular procedures. These may include the code optimization, database optimization (queries, indexing, etc.) or, less often, the transition to more relevant technology or version, introduction of such practices as the transfer of static content to CDN servers, use of distributed infrastructure, and much more.

As a result, your IT system will always be fast and rank higher in search results than those of your competitors.

What you get:

• better resource performance;
• readiness for high loads;
• user retention;
• competitive advantage.

A mobile IT product, that we worked with, had no serious flaws other than poor resilience to a workload with a long response time and limited number of concurrent users. There was no assurance in the system's stability leading to doubts on whether the servers could handle an increase in the load or recover quickly from a system failure.

Our test automation experts checked if the system meets performance requirements. It turned out that the bandwidth was poor, with performance below the required level, while the limits and configuration files had suboptimal settings.

We implemented the comprehensive work to modify the configuration of the hardware and operating system parameters, reconfigure the connection pool and application server, and change indexing settings for database tables.

Following this optimization, the system can now handle more than five thousand concurrent users. The bandwidth increased 15-fold, and the server is responding much faster to key queries. We localized critical resources for further optimization, selected and checked the server configuration for stable operation of the app.

Using Life Hacks for Smooth Operation

Based on our experience, we made recommendations on how to ensure the effective operation of your IT system:

1. Conduct load testing regularly or before an expected increase in traffic. This can help you identify critical points and prepare an optimization plan.
2. Hire an experienced team of DevOps specialists. Implement internal resource monitoring, get regular reports and monitor the load. For example, when the server runs out of space or has performance issues, you will have time to prevent system failure and expand resources.
3. Set up the automation of your order-taking system. This point is important for retail, where the key management decisions are related to improving the quality and speed of customer service. IT systems enable you to eliminate the need for maintaining a large warehouse. For example, when you run out of ingredients, the procurement system will simply send an order to the supplier.

You can learn more about our processes here. For more cases and useful materials for business, visit our LinkedIn and Medium accounts.