How to use the command 'loadtest' (with examples)

The loadtest command is a robust utility designed to perform load testing on web services, specifically HTTP and WebSockets URLs. It assists in determining how a server performs under high demand by simulating multiple users accessing the service simultaneously. This tool is highly beneficial for developers and testers who need to assess the resilience and scalability of web applications before they are deployed in a production environment. One can utilize loadtest to emulate different user loads, customize headers, and employ various HTTP methods, making it a versatile tool for performance evaluation.

Use case 1: Running a Load Test with Concurrent Users and Specified Requests per Second

Code:

loadtest --concurrency 10 --rps 200 https://example.com

Motivation:
When preparing a web service for high-traffic situations, it is crucial to understand how it handles concurrent users sending a specific number of requests per second (RPS). This simulation is especially important for applications expected to receive high volumes of traffic, such as during a major online sale or product launch. By running this load test scenario, developers and testers can identify potential bottlenecks or failures that could arise under stress.

Explanation:

• --concurrency 10: This argument specifies that the load test will simulate 10 concurrent users. Each simulated user will send requests to the server simultaneously, mirroring real-world scenarios where multiple users access the service at the same time.
• --rps 200: It sets the rate of requests per second to 200. With this setting, the tool will dispatch a total of 200 requests every second across the 10 concurrent users.
• https://example.com: This is the target URL where the load test is executed. In real scenarios, this URL should be replaced with the actual service endpoint being tested.

Example Output:

Requests: 5000
Concurrency Level: 10
Request per second: 200
Approx. response time: 300ms

Use case 2: Running a Load Test with a Custom HTTP Header

Code:

loadtest --headers "accept:text/plain;text-html" https://example.com

Motivation:
There are scenarios where the response of a web service varies based on the headers it receives, especially in applications where content negotiation is implemented. By testing a web service with specific HTTP headers, developers can ensure that the service behaves correctly and efficiently for clients that request different content types or other custom headers.

Explanation:

• --headers "accept:text/plain;text-html": This argument sets a custom HTTP header in each request. The accept header indicates that the client expects responses in text/plain or text-html format from the server.
• https://example.com: Similar to previous examples, this is the endpoint where the load test is performed. It should be replaced with the actual service URL under test.

Example Output:

Total requests: 1000
Responses with content type 'text/plain': 700
Responses with content type 'text-html': 300

Use case 3: Running a Load Test with a Specific HTTP Method

Code:

loadtest --method GET https://example.com

Motivation:
Different HTTP methods such as GET, POST, PUT, or DELETE may represent different operational contexts in a web service. Testing these individually ensures each method’s implementation handles load correctly. A GET request is one of the most frequently used HTTP operations to retrieve data. By focusing on this method, developers can verify the server’s ability to efficiently handle read operations under load.

Explanation:

• --method GET: This argument specifies the use of the GET HTTP method in the load test. GET is used here to simulate multiple users attempting to retrieve data from the server, which is a typical scenario in web applications.
• https://example.com: As with previous examples, this is the targeted test URL indicating where the load testing takes place.

Example Output:

Total GET requests: 1000
Average response time: 200ms
Successful responses: 980
Timeout errors: 20

Conclusion:

The loadtest command offers a powerful way to stress-test web applications by simulating multiple users and varied request patterns. Whether you are testing for specific rates of concurrent access, examining the behavior of different HTTP headers, or assessing your application’s response to different HTTP methods, this tool provides critical insights into how your service will perform under pressure. Understanding these aspects allows developers and businesses to optimize their applications for reliability, ensuring that they can withstand the demands of real-world use.