Load testing Python web apps with Gatling

01 🐍 load testing with gatling.md

Results

These results are generated by a gatling simulation that ramps 100 user over 10 second and then keeps the users constant at that rate for 60 seconds before ramping again by 100 users. The entire scenario runs from 0 to 1000 users.

• Overview of Results

Flask

• Werkzeug
• Gunicorn / 1 Workers
• UWSGI / 1 Workers

aiohttp

• aiohttp run_app
• Gunicorn / 1 Workers
• Gunicorn + UVLoop / 1 Workers

02 scenario_example.scala

class flaskGunicornConstantRamp extends Simulation {
  val flaskGunicorn = http.baseURL("http://flask-gunicorn:8080")

  val scn = scenario(
    "flask - gunicorn " +
    "(" + Constants.Workers + " workers ) - " +
    Constants.ConcurrentUsers + " users")
    .exec(http("Endpoint " + Constants.Endpoint).get(Constants.Endpoint))

  val rampTime = 10 seconds
  val runTime = 60 seconds

  setUp(
      scn.inject(
          rampUsersPerSec(0) to(100) during(rampTime),
          constantUsersPerSec(100) during(runTime),
          rampUsersPerSec(100) to(200) during(rampTime),
          constantUsersPerSec(200) during(runTime),
          rampUsersPerSec(200) to(300) during(rampTime),
          constantUsersPerSec(300) during(runTime),
          rampUsersPerSec(300) to(400) during(rampTime),
          constantUsersPerSec(400) during(runTime),
          rampUsersPerSec(400) to(500) during(rampTime),
          constantUsersPerSec(500) during(runTime),
          rampUsersPerSec(500) to(600) during(rampTime),
          constantUsersPerSec(600) during(runTime),
          rampUsersPerSec(600) to(700) during(rampTime),
          constantUsersPerSec(700) during(runTime),
          rampUsersPerSec(700) to(800) during(rampTime),
          constantUsersPerSec(800) during(runTime),
          rampUsersPerSec(800) to(900) during(rampTime),
          constantUsersPerSec(900) during(runTime),
          rampUsersPerSec(900) to(1000) during(rampTime),
          constantUsersPerSec(1000) during(runTime)
      ).protocols(flaskGunicorn)
  )
}

03 server details.json

{
    "ConfigVersion": 3,
    "Driver": {
        "IPAddress": "⋆⋆⋆⋆⋆⋆⋆⋆⋆⋆⋆",
        "MachineName": "gatling-test-server",
        "SSHUser": "root",
        "SSHPort": 22,
        "SSHKeyPath": "~/.docker/machine/machines/gatling-test-server/id_rsa",
        "StorePath": "~/.docker/machine",
        "SwarmMaster": false,
        "SwarmHost": "tcp://0.0.0.0:3376",
        "SwarmDiscovery": "",
        "AccessToken": "⋆⋆⋆⋆⋆⋆⋆⋆",
        "DropletID": "⋆⋆⋆⋆⋆⋆⋆⋆",
        "DropletName": "",
        "Image": "ubuntu-16-04-x64",
        "Region": "tor1",
        "SSHKeyID": "⋆⋆⋆⋆⋆⋆⋆⋆",
        "SSHKeyFingerprint": "",
        "SSHKey": "",
        "Size": "2gb",
        "IPv6": false,
        "Backups": false,
        "PrivateNetworking": false,
        "UserDataFile": ""
    }