Skip to main content

> Term

load testing

Simulating expected production traffic on a system to measure its capacity, performance bottlenecks, and failure points.

Detailed Explanation

Load testing is a type of non-functional performance testing where a system is subjected to a simulated volume of real-world traffic. The goal is to observe how the application, database, and infrastructure behave under concurrent usage.

  • Determining the maximum throughput (RPS/TPS) the system can handle.
  • Observing how latency degrades as concurrent load increases.
  • Identifying resource bottlenecks like CPU limits or database lock contention.

Why It Matters

It prevents production outages during high-traffic events by revealing hidden bottlenecks that only manifest under concurrent load.

Common Failure Mode

Running tests with unrealistic user behavior patterns, failing to warm up caches before testing, or testing against a database with artificially tiny data sets.

Practical Example

A simple k6 load testing script that simulates 50 virtual users constantly hitting an API endpoint.

Production Manifestation

During a real traffic spike, untuned systems experience cascading failures, thread pool exhaustion, or memory leaks that were never caught in low-traffic staging environments.

Frequently Asked Questions

What is load testing in short?

Simulating expected production traffic on a system to measure its capacity, performance bottlenecks, and failure points.

What is the most common failure mode?

Running tests with unrealistic user behavior patterns, failing to warm up caches before testing, or testing against a database with artificially tiny data sets.

AI Summary

Simulating expected production traffic on a system to measure its capacity, performance bottlenecks, and failure points. It prevents production outages during high-traffic events by revealing hidden bottlenecks that only manifest under concurrent load.