Testing Guides

Performance engineering is the practice of building speed, scalability, and stability into software throughout its lifecycle. This guide covers the performance engineering process, the responsibilities involved, the tools teams commonly use, and the role of performance testing.

Load testing (including stress testing, spike testing, and stability testing) is the technique of simulating real users interacting with your site, to see how it performs under heavy load. Proper load testing helps to reduce the risk of site failure in the event of a traffic surge.

We've all seen websites crash under a flood of traffic. When it happens to your site, it's costly and stressful! Preparing your site so it doesn't crash under heavy traffic is often a worthwhile investment, but how can you predict the breaking point at which your site will crash or become slow under heavy load?

Performance testing measures how a website, application, or API behaves under realistic and heavy usage. This guide covers load, stress, soak, and spike testing, the metrics each test produces, and a practical process for planning, running, and analyzing performance tests.

Your customers want to build on an API that's fast and reliable. Performance problems present a serious risk to your API's stability, but a little performance testing in the right places can greatly reduce your risk of API failure. Because they're meant to be consumed programmatically, APIs are usually great candidates for automated, including load and performance testing.

Load testing, stress testing, and DDoS attacks all generate heavy traffic against a web server, so the terms get conflated. This guide explains what each one actually means, who does them and why, and how to tell whether a particular test is legitimate QA or something else.

Synthetic monitoring runs scripted checks against your site or API on a schedule from outside your infrastructure. This guide explains how synthetic monitoring works, how it differs from real user monitoring and load testing, and what to consider when choosing and configuring a monitoring tool.

Load testing and stress testing both simulate heavy traffic on a website, API, or application, but the motivations, methods, and expected outcomes are different. This guide explains the difference, when each is appropriate, and how most teams end up doing both.

Performance tuning web applications isn't easy, but it helps to have a methodical and scientific approach. Tuning gets a lot easier when you have a way to generate consistently reproducible traffic and measure the impact of each and every change. Easier said than done? Here's a quick and dirty approach that gets you most of the way there.

Web applications have more moving parts than static sites — authentication, stateful flows, databases, real-time updates, and client-side rendering all interact under load. Stress testing a web application means pushing it past capacity while watching these components for degradation, failure, and recovery.

We make a load testing tool, so it would be tempting to say that you should load test your site early and often, extensively, all the time, no matter what! But the real answer is: it depends. Sometimes it isn't worth it. There are hard and soft costs to load testing that you should weigh against the likelihood and severity of the risk you're trying to mitigate.

Capacity planning means making sure your servers, cloud infrastructure, and web applications have the capacity to service the expected demand, allowing for fluctuations and traffic spikes. But how can you get the most bang for your buck when capacity planning, without it turning into a dull and cumbersome process? Here is a short, opinionated crash course.

Every website has a concurrency limit, but server specifications alone can't tell you what it is. This guide gives rough ranges for common hosting setups, explains where bottlenecks typically appear, and shows how to measure your actual limit with a stress test.

Capacity management is the ongoing practice of making sure your systems have enough capacity to meet demand at a cost within your budget. This guide covers the capacity management cycle, how it differs from capacity planning, and the role of load testing and monitoring.

Concurrent users are the number of visitors actively using your site at the same time. This guide explains how concurrency differs from visits, sessions, and requests per second, and how to estimate concurrent users from your analytics data for capacity planning and load testing.

Soak testing (also called endurance testing or stability testing) applies steady, moderate load to an application over a long stretch of time (hours or days) to surface problems that don't appear in short tests. This guide covers what soak testing is, what it typically finds, how to run one, and how it fits alongside load and stress testing.

Is the adage true, that 80% of the time that end users spend waiting on a web page is due to front-end factors? Maybe under normal conditions. But under peak traffic, that ratio can flip. Here's a bit of perspective on where to focus your efforts to ensure your site stays fast and delivers a consistent experience to all your users, all the time.

Concurrency testing verifies that a system behaves correctly when many users perform actions at the same time. This guide covers common concurrency bugs, how to create simultaneous traffic with load testing bots, and how to verify the resulting application data.

Capacity testing measures the maximum load a system can handle while still meeting its response time and error-rate targets. This guide explains how capacity testing differs from stress testing, how to run a capacity test, and how to use the results for infrastructure and traffic planning.

Scalability testing measures how a system's performance changes as load and infrastructure grow. This guide covers scaling up versus scaling out, how to compare configurations with stepped load tests, and how to use the results for capacity and infrastructure planning.

Spike testing simulates a short, intense burst of traffic to measure how your web application performs during the surge and whether it safely recovers afterwards. This guide explains how spike testing differs from a standard load test, how to build a realistic spike scenario, and how to interpret results.

Load testing isn't always easy, but doing it right can greatly reduce your risk of site crashes. We've worked with thousands of testers and performance engineers, and assembled a list of load testing best practices and pitfalls to help you be successful in less time.