Locust
Python-based open-source load testing with user behavior defined in code.
About
Locust is an open-source load testing framework where you describe simulated user behavior in plain Python and it runs thousands of those users concurrently, reporting requests per second, latency percentiles, and failure rates live through a web UI or headless CI mode.
It has become a standard choice for load testing AI applications: teams use it to hammer agent chat endpoints with weighted conversation tasks, measure time to first token and turn latency against Vertex AI Agent Engine deployments, and wire SLO assertions into staging pipelines. Distributed runs scale across workers, and its FastHttpUser mode delivers much higher throughput per machine.
What it can do
- User behavior defined in plain Python
- Live web UI with percentiles and failure rates
- Headless mode for CI pipelines
- Native distributed load generation
- FastHttpUser for high-throughput scenarios
Setup
pip install locust, write a locustfile.py with user classes, then run locust -f locustfile.py --host <target>.
Prerequisites
Python; a target application or API endpoint to test.
Use cases
Works with
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 90; Trending score 72/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.