Locust
Open-source load testing tool where you write test scenarios as plain Python code.
- Stars
- 28.3k
- Forks
- 3.3k
- Licence
- MIT
- Language
- Python
- Last update
- 4 days ago
- Type
- App
What it is
Locust is a free, open-source tool for load testing, which means simulating many users at once to see how a web service or other system behaves under pressure. You describe user behavior in ordinary Python code, then run the test from the command line or a web UI that shows throughput, response times, and errors in real time. Its README says it works mainly with HTTP but can be extended to other protocols.
- Simulate a login flow followed by repeated API requests, with weighted task selection
- Run the same test headless in a CI pipeline and fail the build on errors
- Change the number of simulated users while a test is running through the web UI
- Distribute a test across multiple machines using its worker mode
- Export test results for later analysis
Great for
- Backend or QA engineers who want load tests kept in version control as normal code
- Python teams who want to reuse existing libraries inside their tests
- Teams running load tests in CI/CD pipelines without a GUI
- Anyone who needs to test highly concurrent workloads from several machines
Not ideal for
- People who want a no-code tool, since tests are written in Python
- Teams testing non-HTTP systems who are unwilling to write a custom client, since the README says other protocols need one
- Users looking for a hosted service, since the README points to Locust Cloud and Azure options separately rather than describing a built-in hosted offering
How to get it
- 1
Install with a package manager
You'll need: Python
Install from PyPI
pip install locust - 2
Run it
- Install Locust with pip
- Write a test file that defines a user class with tasks, such as the QuickstartUser example in the README
- Start Locust and open the web UI, or run it headless from the command line
- Set the number of users and target host, then start the test and watch the charts
- 3
Get the most out of it
- Use custom load shapes to control how user counts change over time, as covered in the docs
- Send real-time reporting data to TimescaleDB and Grafana using community plugins listed in the README
- Scale out by running worker machines for distributed tests
- Explore the community extensions listed on the project wiki
- Use the VS Code extension from Microsoft if you want Copilot-assisted test creation
Downloads
Strengths
- Tests are plain Python, so they work with IDEs, version control, and existing libraries
- Each simulated user is lightweight, so one process can handle many concurrent users according to the README
- Offers both a web UI and a headless mode for automation
Trade-offs
- Requires Python knowledge to write tests, and the README describes it as primarily aimed at web services
- Performance per machine depends on hardware; the README acknowledges other tools may handle more requests per second on the same hardware
- Testing non-HTTP systems requires writing your own client code
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| locust (this) | 28.3k | App | intermediate | 4 days ago | Backend or QA engineers who want load tests kept in version control as normal code |
| vegeta | 25.2k | CLI | intermediate | 1 months ago | Backend developers checking how an HTTP API behaves under a fixed request rate |
| oha | 10.7k | CLI | beginner | 4 days ago | Developers checking how an API or website handles many simultaneous requests |
| systeminformer | 16.4k | App | intermediate | 4 days ago | Windows users who want more detail than Task Manager provides |
| BenchmarkDotNet | 11.5k | Library | intermediate | 5 days ago | .NET developers who want to compare two or more implementations of a method |
| nginx-admins-handbook | 14.3k | Learning resource | beginner | 2.0 years ago | System administrators who already run NGINX and want practical tips |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.