What Is Load Testing?

What Load Testing a Website Involves

Load testing website performance describes sending simulated traffic to a site to see how it behaves as demand grows. A load test shows whether response times stay acceptable, whether errors appear and which part of the system runs out of capacity first.

A speed test measures how one visitor experiences a page. Load testing asks a different question: what happens when many visitors arrive together? A page that loads in two seconds for one person may respond far more slowly when hundreds of people request it at the same time.

Load, Stress, Spike and Soak Tests

Testers use different traffic patterns to answer different questions:

• Load test checks behavior at the expected busy period, such as a normal peak hour.
• Stress test pushes traffic beyond expected levels to find the breaking point and see how the system recovers.
• Spike test adds a sudden surge, such as the traffic after a mention on television or a promotional email, to see whether the site copes with sharp changes.
• Soak test holds a moderate load for many hours, which can reveal memory leaks or resources that slowly run out.

Each type answers a different question, so a site that passes a normal load test may still fail under a spike.

Concurrent Users vs Requests per Second

Two numbers are often confused. Concurrent users are people active on the site at the same moment. Requests per second counts how many HTTP requests the server receives each second.

The two are related but not equal. A visitor who loads a page, reads for a minute and then clicks again creates a burst of requests followed by silence. So 1,000 concurrent users may produce far fewer requests per second than 1,000 users clicking constantly. Requests per second is usually the more useful measure of server capacity, because it reflects the work the server actually does.

Reading Results Without Being Misled

Averages can hide problems. A test may show a healthy average response time while a meaningful share of visitors waits many seconds. Percentile figures, such as the 95th percentile, show what the slowest visitors experience and are more informative.

Look for the point where response times rise sharply. A gradual increase at higher load is normal, but a steep curve often means a limit has been reached, such as available PHP workers, database connections or memory. Error rates and the component that saturates first are equally important.

Results also depend on the setup. A test run against a different environment, with less data or without real third-party services, may not predict production behavior. The testing tool itself can become a bottleneck if it runs on a weak machine. Load tests should only be run against systems you own or have permission to test, because heavy traffic can disrupt a live service.