Load testing 1-on-1: discovering the limits of your system

Benchmarking is hard. Ever wondered how reliable is your system? How many requests per second can it handle? How about the average response time: what do you do to measure that? How does the system behave during events like black friday, christmas sales, or any other important date that may cause an outage? Have no idea where to start answering these questions? Fear not, because you are in the right place!

In this tutorial I will introduce you to locust: probably the most awesome pythonic tool to perform load tests against websites, REST APIs, GQL, and many other ways of contacting and using systems. You'll be able to understand how to structure your stress tests projects, and get your hands dirty in creating tests against a REST API, while visually being able to tell where is the threshold point and what happens after that.

This tutorial is targeted for intermediate Pythonistas. Some recommended requirements (or things that will not be covered in a greater level of details) include: basics of git, simple Docker commands (to get a container up and running), classes in Python, some understanding of REST APIs. Will do my best to make it possible so anyone/any level of knowledge can join. After this presentation, participants will be able to create tests that benchmark applications in the most realistic way: exactly how users would use a system - and with that, answer the initial questions proposed, as well as create reports to display anywhere necessary (such as your job).