Introduction

Spike testing is a type of performance testing that evaluates how a system responds to sudden, extreme increases in load. Unlike stress testing, which gradually increases the load, spike testing simulates abrupt surges in traffic to identify system vulnerabilities, such as crashes, slow response times, and resource exhaustion.

In this blog, we will explore spike testing in detail, covering its importance, methodology, and full implementation using K6.

Why Perform Spike Testing?

Spike testing helps you

• Determine system stability under unexpected traffic surges.
• Identify bottlenecks that arise due to rapid load increases.
• Assess auto-scaling capabilities of cloud-based infrastructures.
• Measure response time degradation during high-demand spikes.
• Ensure system recovery after the sudden load disappears.

Setting Up K6 for Spike Testing

Installing K6

# macOS
brew install k6  

# Ubuntu/Debian
sudo apt install k6  

# Using Docker
docker pull grafana/k6  

Choosing the Right Test Scenario

K6 provides different executors to simulate load patterns. For spike testing, we use

• ramping-arrival-rate → Gradually increases the request rate over time.
• constant-arrival-rate → Maintains a fixed number of requests per second after the spike.

Example 1: Basic Spike Test

This test starts with low traffic, spikes suddenly, and then drops back to normal.

import http from 'k6/http';
import { sleep } from 'k6';

export let options = {
  scenarios: {
    spike_test: {
      executor: 'ramping-arrival-rate',
      startRate: 10, // Start with 10 requests/sec
      timeUnit: '1s',
      preAllocatedVUs: 100,
      maxVUs: 500,
      stages: [
        { duration: '30s', target: 10 },  // Low traffic
        { duration: '10s', target: 500 }, // Sudden spike
        { duration: '30s', target: 10 },  // Traffic drops
      ],
    },
  },
};

export default function () {
  http.get('https://test-api.example.com');
  sleep(1);
}

Explanation

• Starts with 10 requests per second for 30 seconds.
• Spikes to 500 requests per second in 10 seconds.
• Drops back to 10 requests per second.
• Tests the system’s ability to handle and recover from traffic spikes.

Example 2: Spike Test with High User Load

This test simulates a spike in virtual users rather than just requests per second.

import http from 'k6/http';
import { sleep } from 'k6';

export let options = {
  scenarios: {
    user_spike: {
      executor: 'ramping-vus',
      stages: [
        { duration: '30s', target: 20 },  // Normal traffic
        { duration: '10s', target: 300 }, // Sudden spike in users
        { duration: '30s', target: 20 },  // Drop back to normal
      ],
    },
  },
};

export default function () {
  http.get('https://test-api.example.com');
  sleep(1);
}

Explanation:

• Simulates a sudden increase in concurrent virtual users (VUs).
• Helps test server stability, database handling, and auto-scaling.

Example 3: Spike Test on Multiple Endpoints

In real-world applications, multiple endpoints may experience spikes simultaneously. Here’s how to test different API routes.

import http from 'k6/http';
import { sleep } from 'k6';

export let options = {
  scenarios: {
    multiple_endpoint_spike: {
      executor: 'ramping-arrival-rate',
      startRate: 5,
      timeUnit: '1s',
      preAllocatedVUs: 200,
      maxVUs: 500,
      stages: [
        { duration: '20s', target: 10 },  // Normal traffic
        { duration: '10s', target: 300 }, // Spike across endpoints
        { duration: '20s', target: 10 },  // Traffic drop
      ],
    },
  },
};

export default function () {
  let urls = [
    'https://test-api.example.com/users',
    'https://test-api.example.com/orders',
    'https://test-api.example.com/products'
  ];
  
  let res = http.get(urls[Math.floor(Math.random() * urls.length)]);
  console.log(`Response time: ${res.timings.duration}ms`);
  sleep(1);
}

Explanation

• Simulates traffic spikes across multiple API endpoints.
• Helps identify which API calls suffer under extreme load.

Analyzing Test Results

After running the tests, K6 provides key performance metrics

http_req_duration......: avg=350ms min=150ms max=3000ms
http_reqs..............: 10,000 requests
vus_max................: 500
errors.................: 2%

Key Metrics

• http_req_duration → Measures response time impact.
• vus_max → Peak virtual users during the spike.
• errors → Percentage of failed requests due to overload.

Best Practices for Spike Testing

• Monitor application logs and database performance during the test.
• Use auto-scaling mechanisms for cloud-based environments.
• Combine spike tests with stress testing for better insights.
• Analyze error rates and recovery time to ensure system stability.

Spike testing is crucial for ensuring application stability under sudden, unpredictable traffic surges. Using K6, we can simulate spikes in both requests per second and concurrent users to identify bottlenecks before they impact real users.

Introduction

Stress testing is a critical aspect of performance testing that evaluates how a system performs under extreme loads. Unlike load testing, which simulates expected user traffic, stress testing pushes a system beyond its limits to identify breaking points and measure recovery capabilities.

In this blog, we will explore stress testing using K6, an open-source load testing tool, with detailed explanations and full examples to help you implement stress testing effectively.

Why Stress Testing?

Stress testing helps you

• Identify the maximum capacity of your system.
• Detect potential failures and bottlenecks.
• Measure system stability and recovery under high loads.
• Ensure infrastructure can handle unexpected spikes in traffic.

Setting Up K6 for Stress Testing

Installing K6

# macOS
brew install k6  

# Ubuntu/Debian
sudo apt install k6  

# Using Docker
docker pull grafana/k6  

Understanding Stress Testing Scenarios

K6 provides various executors to simulate different traffic patterns. For stress testing, we mainly use

1. ramping-vus – Gradually increases virtual users to a high level.
2. constant-vus – Maintains a fixed high number of virtual users.
3. spike – Simulates a sudden surge in traffic.

Example 1: Basic Stress Test with Ramping VUs

This script gradually increases the number of virtual users, holds a peak load, and then reduces it.

import http from 'k6/http';
import { sleep } from 'k6';

export let options = {
  stages: [
    { duration: '1m', target: 100 }, // Ramp up to 100 users in 1 min
    { duration: '3m', target: 100 }, // Stay at 100 users for 3 min
    { duration: '1m', target: 0 },   // Ramp down to 0 users
  ],
};

export default function () {
  let res = http.get('https://test-api.example.com');
  sleep(1);
}

Explanation

• The test starts with 0 users and ramps up to 100 users in 1 minute.
• Holds 100 users for 3 minutes.
• Gradually reduces load to 0 users.
• The sleep(1) function helps simulate real user behavior between requests.

Example 2: Constant High Load Test

This test maintains a consistently high number of virtual users.

import http from 'k6/http';
import { sleep } from 'k6';

export let options = {
  vus: 200, // 200 virtual users
  duration: '5m', // Run the test for 5 minutes
};

export default function () {
  http.get('https://test-api.example.com');
  sleep(1);
}

Explanation

• 200 virtual users are constantly hitting the endpoint for 5 minutes.
• Helps evaluate system performance under sustained high traffic.

Example 3: Spike Testing (Sudden Traffic Surge)

This test simulates a sudden spike in traffic, followed by a drop.

import http from 'k6/http';
import { sleep } from 'k6';

export let options = {
  stages: [
    { duration: '10s', target: 10 },  // Start with 10 users
    { duration: '10s', target: 500 }, // Spike to 500 users
    { duration: '10s', target: 10 },  // Drop back to 10 users
  ],
};

export default function () {
  http.get('https://test-api.example.com');
  sleep(1);
}

Explanation

• Starts with 10 users.
• Spikes suddenly to 500 users in 10 seconds.
• Drops back to 10 users.
• Helps determine how the system handles sudden surges in traffic.

Analyzing Test Results

After running the tests, K6 provides detailed statistics

checks..................: 100.00% ✓ 5000 ✗ 0
http_req_duration......: avg=300ms min=200ms max=2000ms
http_reqs..............: 5000 requests
vus_max................: 500

Key Metrics to Analyze

• http_req_duration → Measures response time.
• vus_max → Maximum concurrent virtual users.
• http_reqs → Total number of requests.
• errors → Number of failed requests.

Stress testing is vital to ensure application stability and scalability. Using K6, we can simulate different stress scenarios like ramping load, constant high load, and spikes to identify system weaknesses before they affect users.