I'm evaluating different approaches to data access and other cross-cutting concerns in .NET 7 Web APIs, both for clients and in preparation for an upcoming conference talk. I wanted to be able to load test endpoints locally as one way to assess performance. Looking for good local tools that aren't bloated and hard to use (yes I'm talking to you, JMeter), I quickly found K6. It seemed to serve my needs pretty well for the local web-based load testing need I had. However, I also wanted to be able to drill down into individual methods within certain endpoints, so I found myself also using Benchmark.NET.

Getting started with K6 was quite easy:

choco install k6

Run scripts from the k6 folder.

k6 run simple.js

An example script (simple.js):

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

export default function () {
  http.get('https://test.k6.io');
  sleep(1);
}

// usage
// k6 run simple.js
// k6 run --vus 10 --duration 30s simple.js

Leveraging AI

At the same time, I'm also evaluating how ChatGPT can help make me more productive, so I asked it:

When would I use the tool k6 for load or performance testing my .NET web api application, and when would I use Benchmark.NET tool for the same thing?

Of course, I have my own opinion about this, but its response was actually pretty good:

k6 is a load testing tool that can be used to measure the performance and scalability of your .NET web API application. It supports writing tests in JavaScript and can be used to simulate a large number of virtual users making requests to your API. k6 is best suited for load testing scenarios where you want to simulate a large number of concurrent users and measure the performance of your API under load.

On the other hand, Benchmark.NET is a microbenchmarking tool that can be used to measure the performance of individual methods or pieces of code within your .NET application. It is best suited for performance testing scenarios where you want to measure the performance of specific pieces of code in isolation.

In summary, you would use k6 for load testing your .NET web API application to simulate a large number of concurrent users and measure the overall performance and scalability of your API under load. You would use Benchmark.NET to measure the performance of specific pieces of code within your application in isolation.

Of course, K6 and Benchmark.NET aren't the only tools one might use for this purpose. I didn't want to get into cloud-based testing tools like BlazeMeter, so I narrowed my question a bit for this next prompt. And remember, this version of ChatGPT only knows about things circa 2021. I asked:

What other options should I consider to check the performance of my .NET web api project? Note I'm only interested in free tools I can use locally on my computer.

And ChatGPT responded:

Here are some other free tools you can consider to check the performance of your .NET web API project:

• Apache JMeter: Apache JMeter is a free and open-source tool that can be used for load testing your .NET web API application. It supports multiple protocols including HTTP, HTTPS, and FTP, and can be used to simulate a large number of concurrent users.

• Gatling: Gatling is another open-source load testing tool that can be used to test the performance of your .NET web API application. It supports scripting in Scala and can simulate a large number of concurrent users making requests to your API.

• LoadUI: LoadUI is a free and open-source load testing tool that can be used to test the performance of your .NET web API application. It supports multiple protocols including HTTP, HTTPS, and SOAP, and can simulate a large number of virtual users.

• Siege: Siege is a free and open-source tool that can be used for load testing your .NET web API application. It supports HTTP and HTTPS protocols and can simulate a large number of concurrent users.

• Apache Bench (ab): Apache Bench is a command-line tool that can be used to test the performance of your .NET web API application. It is a simple tool that can be used to send HTTP requests to your API and measure the response time.

Note that while these tools are free to use, they may have limitations in terms of the number of virtual users or requests they can simulate. It is important to check the documentation and understand the limitations of each tool before deciding to use it for your load testing needs.

Summary

Overall I'm pleased with the information ChatGPT was able to provide - it certainly was more comprehensive and useful than if I'd attempted a Google search. And, get this, there were no "sponsored links" or other ads involved! How amazing is that? It also helped me discover some resources that I didn't know I didn't know about, such as LoadUI and Siege, which I'll need to check out further.

References

• Apache JMeter
• Gatling
• LoadUI
• Siege
• Apache Bench (ab)
• k6
• Benchmark.NET
• .NET web API documentation
• Load testing best practices
• Performance testing with Visual Studio
• Testing ASP.NET Core Web API with XUnit