Run your first test
Use Docker Compose and make quickstart to hit the bundled mock target with zero
manual worker provisioning.
test-scale documentation
Distributed HTTP load testing — YAML in, autoscaling workers, Prometheus + Grafana out.
What you will learn
Section titled “What you will learn”Use the web UI
Submit plans, watch runs, compare baselines, and open Grafana from the embedded control-plane UI — see the Web UI guide.
Test real APIs
Point workers at host or remote URLs, send JSON bodies, and authenticate with
Bearer/JWT or session cookies — secrets via ${ENV_VAR} in plan files.
Trust the numbers
Understand when local Compose results are enough for regression testing and when to move to Kubernetes for scale-out load generation.
How test-scale works (30 seconds)
Section titled “How test-scale works (30 seconds)”- You submit a test plan — a small YAML file describing one HTTP request and how hard to hit it.
- The control plane validates the plan, computes how many workers you need, and asks the autoscaler to start them.
- Workers pull work from NATS, generate load against your target, and stream metrics to the aggregator.
- Prometheus and Grafana show live throughput and latency; every run is stored so you can compare against a baseline.
test-scale is built for repeatable performance regression testing, not for multi-step browser journeys or chaos engineering.