k6 is an open-source load testing tool designed for modern developers and engineers. It is simple to use, yet powerful, enabling the creation, execution, and analysis of performance tests. With k6, you can easily simulate a large number of virtual users, define complex test scenarios, and collect performance metrics to identify bottlenecks and optimize your applications.
- Scripting in JavaScript: Write test scenarios using a JavaScript-based scripting language.
- Command-Line Interface (CLI): Run tests from the command line, integrate with CI/CD pipelines.
- Built-in Metrics: Collects detailed metrics on response times, request rates, and more.
- Extensible: Supports plugins and extensions to enhance functionality.
- Local and Cloud Execution: Run tests locally or in the cloud with k6 Cloud.
- Node.js: Ensure Node.js is installed on your machine.
- Homebrew (macOS): For easy installation on macOS.
brew install k6docker pull loadimpact/k6Download the latest release from the official k6 GitHub repository.
Create a JavaScript file (e.g., test.js) with the following content:
import http from 'k6/http';
import { sleep, check } from 'k6';
export let options = {
stages: [
{ duration: '30s', target: 20 },
{ duration: '1m', target: 10 },
{ duration: '10s', target: 0 },
],
};
export default function () {
let res = http.get('https://test.k6.io');
check(res, {
'is status 200': (r) => r.status === 200,
});
sleep(1);
}Run the script using the k6 CLI:
k6 run test.jsAfter running the test, k6 will provide a summary of the test results, including:
- HTTP request metrics: Response time, request count, failure rate.
- System metrics: CPU and memory usage.
- Custom metrics: Defined within your scripts.
Customize your load tests with various options:
- Thresholds: Define performance thresholds to automatically fail tests if they are not met.
- Tags: Add tags to requests and metrics for better organization and analysis.
- Scenarios: Define complex user behavior and traffic patterns.
k6 can be integrated into your CI/CD pipelines for automated performance testing. Use tools like Jenkins, GitLab CI, GitHub Actions, and more.
Leverage the power of the cloud for distributed load testing with k6 Cloud. Seamlessly scale your tests and analyze results with advanced features.
- Documentation: Official k6 Documentation
- GitHub: k6 GitHub Repository
- Community Forum: k6 Community Forum
- Slack Channel: Join the k6 Slack channel for support and discussions.
k6 is a versatile and powerful tool for load testing, designed to meet the needs of modern development practices. Its simplicity, combined with robust features, makes it an excellent choice for developers and engineers looking to ensure the performance and reliability of their applications.
/\ |‾‾| /‾‾/ /‾‾/
/\ / \ | |/ / / /
/ \/ \ | ( / ‾‾\
/ \ | |\ \ | (‾) |
/ __________ \ |__| \__\ \_____/ .io
execution: local
script: tests/dashboard/load-test.js
output: -
scenarios: (100.00%) 1 scenario, 100 max VUs, 1m30s max duration (incl. graceful stop):
* default: Up to 100 looping VUs for 1m0s over 3 stages (gracefulRampDown: 30s, gracefulStop: 30s)
✗ is status 200
↳ 70% — ✓ 2030 / ✗ 849
checks.........................: 70.51% ✓ 2030 ✗ 849
data_received..................: 94 MB 1.1 MB/s
data_sent......................: 502 kB 5.9 kB/s
http_req_blocked...............: avg=15.64ms min=100ns med=200ns max=3.28s p(90)=300ns p(95)=409ns
http_req_connecting............: avg=8.33ms min=0s med=0s max=3.21s p(90)=0s p(95)=0s
✗ http_req_duration..............: avg=571.63ms min=24.16ms med=429.67ms max=8.32s p(90)=1.09s p(95)=1.34s
{ expected_response:true }...: avg=689.71ms min=70.79ms med=604.49ms max=8.32s p(90)=1.18s p(95)=1.47s
✗ http_req_failed................: 29.48% ✓ 849 ✗ 2030
http_req_receiving.............: avg=406.67ms min=31.3µs med=301.9ms max=7.24s p(90)=888.3ms p(95)=1.08s
http_req_sending...............: avg=33.03µs min=10.6µs med=28.7µs max=3.27ms p(90)=46.6µs p(95)=62.61µs
http_req_tls_handshaking.......: avg=7.26ms min=0s med=0s max=407.07ms p(90)=0s p(95)=0s
http_req_waiting...............: avg=164.92ms min=21.82ms med=97.1ms max=8.05s p(90)=340.36ms p(95)=613.69ms
http_reqs......................: 2879 34.090558/s
iteration_duration.............: avg=1.58s min=1.02s med=1.44s max=9.33s p(90)=2.11s p(95)=2.37s
iterations.....................: 2879 34.090558/s
vus............................: 1 min=1 max=100
vus_max........................: 100 min=100 max=100
This performance test was executed using k6, a modern load testing tool. The script used for this test is tests/dashboard/load-test.js, and it was run locally. The test scenario aimed to simulate up to 100 virtual users (VUs) over a period of 1 minute with a 30-second ramp-up and ramp-down period, including a graceful stop.
- Maximum VUs: 100
- Test Duration: 1 minute (plus 30s ramp-up and 30s ramp-down)
- Test Script:
tests/dashboard/load-test.js - Load Pattern:
- Ramp up to 100 VUs over 15 seconds
- Hold at 100 VUs for 30 seconds
- Ramp down to 0 VUs over 15 seconds
- Total Requests: 2879
- Requests per Second: 34.09
- Data Received: 94 MB (1.1 MB/s)
- Data Sent: 502 kB (5.9 kB/s)
- is status 200:
- Passed: 2030 (70%)
- Failed: 849 (30%)
-
http_req_blocked:
- Average: 15.64 ms
- Median: 200 ns
- 90th Percentile: 300 ns
- 95th Percentile: 409 ns
- Min: 100 ns
- Max: 3.28 s
-
http_req_connecting:
- Average: 8.33 ms
- Median: 0 s
- 90th Percentile: 0 s
- 95th Percentile: 0 s
- Min: 0 s
- Max: 3.21 s
-
http_req_duration:
- Average: 571.63 ms
- Median: 429.67 ms
- 90th Percentile: 1.09 s
- 95th Percentile: 1.34 s
- Min: 24.16 ms
- Max: 8.32 s
-
http_req_failed:
- Failed Requests: 29.48%
- Passed Requests: 70.52%
-
http_req_receiving:
- Average: 406.67 ms
- Median: 301.9 ms
- 90th Percentile: 888.3 ms
- 95th Percentile: 1.08 s
- Min: 31.3 µs
- Max: 7.24 s
-
http_req_sending:
- Average: 33.03 µs
- Median: 28.7 µs
- 90th Percentile: 46.6 µs
- 95th Percentile: 62.61 µs
- Min: 10.6 µs
- Max: 3.27 ms
-
http_req_tls_handshaking:
- Average: 7.26 ms
- Median: 0 s
- 90th Percentile: 0 s
- 95th Percentile: 0 s
- Min: 0 s
- Max: 407.07 ms
-
http_req_waiting:
- Average: 164.92 ms
- Median: 97.1 ms
- 90th Percentile: 340.36 ms
- 95th Percentile: 613.69 ms
- Min: 21.82 ms
- Max: 8.05 s
-
Iteration Duration:
- Average: 1.58 s
- Median: 1.44 s
- 90th Percentile: 2.11 s
- 95th Percentile: 2.37 s
- Min: 1.02 s
- Max: 9.33 s
-
Iterations: 2879 (34.09 iterations/s)
-
Virtual Users:
- Minimum: 1
- Maximum: 100
1. Response Status Checks
- is status 200: Only 70% of the requests returned a status code of 200, indicating a significant number of requests (30%) failed. This needs to be investigated further to determine the cause of these failures.
2. HTTP Request Duration
- The average request duration was 571.63 ms, with the 95th percentile at 1.34 seconds, which exceeds the desired threshold of 500 ms. This suggests that the system is experiencing performance bottlenecks under load. Potential areas to investigate include:
- Server-side processing times.
- Database performance.
- Network latency.
3. HTTP Request Failures
- The failure rate of HTTP requests was 29.48%, significantly higher than the acceptable threshold of 1%. This high failure rate is concerning and requires immediate attention. Possible causes could include:
- Insufficient server capacity to handle the load.
- Application-level errors or timeouts.
- Network issues.
4. Data Transfer
- The amount of data received and sent seems reasonable given the number of requests. However, the high request durations and failure rates indicate underlying issues affecting performance.
5. Recommendations
- Optimize Backend Performance: Investigate and optimize server-side processing, database queries, and application logic to reduce response times.
- Scale Resources: Consider scaling the server resources (CPU, memory, etc.) to handle higher loads more efficiently.
- Improve Error Handling: Implement better error handling and logging to identify and resolve issues causing request failures.
- Load Balancing: Utilize load balancing techniques to distribute the load more evenly across multiple servers.
Feel free to customize the content further based on your specific requirements or preferences.