Exploring Modern Load Testing with k6 – A QA’s Practical Journey

introduction

While many testers today explore AI tools, I have focused on understanding how applications behave under real-world load conditions.

As a QA engineer, my focus was mostly on functional testing. But an important question always remained in my mind:

• What happens when multiple users access the system at the same time?
• Will the app still respond quickly?
• Or will it fail under pressure?

To answer these questions, I started working with K6a modern and lightweight tool for pregnancy testing.

Why load testing matters

Functional test answers:
Does the feature work?

But performance testing answers:
Will it still work under load?

In real-world systems, performance directly impacts user experience.
This is why a pregnancy test is not optional, it is necessary.

Getting started with K6

I started with a simple text:

import { check, sleep } from 'k6';

export const options = {
  vus: 50,        // 50 virtual users
  duration: '15s', // run for 15 seconds
};

export default function () {
    // Mock Add Employee request
    const res = { status: 201 }; // simulate success
    check(res, { 'employee added (mock)': (r) => r.status === 201 });
    sleep(1);
}

With just a few lines, I was able to simulate multiple users accessing an application simultaneously.

Run the test

Tests running k6.js

k6 runs directly from the command line and immediately starts generating the download.

Understand the results

K6 provides key performance metrics:

• Response time (p. 95, p. 99) → Real user experience
• Error rate → Failure during implementation
• Requests per second (RPS) → Productivity
• Virtual Users (VUs) → Loading simulation

These metrics help evaluate the stability and scalability of the application.

The real challenges I faced

Demo application limitations

While testing beta applications:

• Some actions are blocked
• APIs behaved inconsistently

Learning: Experimental environments are not always reliable.

Testing UI against API

UI-based load testing was:

Switch to testing API provided:

• Faster execution
• More accurate results

Learn: Testing at the API level is more reliable.

Understanding Metrics (Initial Struggle)

At first, metrics like P95 and P99 were confusing.

Later I understood:

The P95’s response time reflects the real user experience better than average

Learning: Understanding metrics is just as important as taking tests.

k6 vs Apache JMeter – practical comparison

My opinion:

• K6 is ideal for modern API testing and automation
• JMeter is useful for GUI-based tests and legacy tests

Artificial intelligence + pregnancy test

I’ve also explored AI-based testing tools that can:

• Generate test cases
• Simulate user flows

but:

AI tools are useful, but strong fundamentals provide better control and accuracy.

Key takeaways

• Start small and expand gradually
• Prefer API testing over user interface (UI) testing.
• Understand metrics before measuring
• Focus on real world scenarios
• Basics > Tools

Rethink testing with K6

This experience changed the way I approached testing.
The test is not only:

Does it work?

But also:

Will he still work under pressure?

Working with k6 helped me understand Performance testing From a practical and confidence-giving perspective to analyze system behavior in the real world.