Artillery is a modern, powerful, easy-to-use load-testing tool.

Artillery has a strong focus on developer happiness & ease of use, and a batteries-included philosophy.

Artillery's goal is to help developers build faster, more resilient and more scalable applications.

Features

Mulitple protocols: Load-test HTTP, WebSocket and Socket.io applications
Scenarios: Specify scenarios to test multi-step interactions in your API or web app
Perfomance metrics: get detailed performance metrics (latency, requests per second, concurrency, throughput)
Scriptable: write custom logic in JS to do pretty much anything
High performance: generate serious load on modest hardware
Integrations: statsd support out of the box for real-time reporting (integrate with Datadog, Librato, InfluxDB etc)
Extensible: custom reporting plugins, custom protocol engines etc
and more! HTML reports, nice CLI, parameterization with CSV files

Source: https://github.com/shoreditch-ops/artillery
Issues: https://github.com/shoreditch-ops/artillery/issues
Chat: https://gitter.im/shoreditch-ops/artillery
Docs: https://artillery.io/docs/
Website: https://artillery.io
Twitter: @ShoreditchOps
Support SLAs, custom integrations, professional services: enterprise@artillery.io

Use Cases

Benchmark the performance of your API or microservice as you're building it
Ensure new code does not introduce performance regressions
Test your code for memory leaks
Benchmark dependencies (libraries, frameworks, external services) to get a feel for their performance characteristics before you integrate
Run load-tests before you launch to ensure your application can meet projected demand

Quickstart

Install

Artillery is available via npm

$ npm install -g artillery

Node.js v4+ is required (Node.js 6 is recommended).

Run A Quick Test

$ artillery quick -d 30 -r 5 -n 20 http://127.0.0.1:3000/test

This will run a test for 30 seconds, with 5 new virtual users created every second, with each user sending 20 a GET requests to http://127.0.0.1:3000/test.

Run With A Scenario

Artillery's power lies in emulating complex behavior, like that of users of an e-commerce website, a transactional API etc.

Run a scenario with:

$ artillery run hello.yaml

Where hello.yaml is your tests script that contains something like:

(NB: test scripts can be written as JSON too)

config:
  target: "http://127.0.0.1:3000"
  phases:
    - duration: 120
      arrivalRate: 10
  defaults:
    headers:
      content-type: "application/json"
      x-my-service-auth: fedcba9876543210
scenarios:
  - flow:
      - get:
          url: "/test"
      - think: 1
      - post:
          url: "/test"
          json:
            name: "Hassy"

This will run a test for 2 minutes, with 10 virtual users created every second, each of which will send a GET and a POST request with a pause of 1 second in between. Each request will include two custom headers (Content-Type and X-My-Service-Auth).

Create A HTML Report

Once the test completes, you can create a graphical report from the JSON stats produced by artillery run with: artillery report <report_xxxxx.json>

These are self-contained HTML files that can be easily shared via email or Dropbox for example.

Learn More

See Artillery docs for docs and examples.

Contributing

Thinking of contributing to Artillery? Awesome! Please have a quick look at the guide.

Using Artillery?

Are you using Artillery to ship faster, more resilient and more scalable systems? Add your team to the Artillery users list on the wiki.

License

Artillery is open-source software distributed under the terms of the MPL2 license.

MPL 2.0 FAQ

Muscot/artillery