The BrowserUp Config Yaml Format

In the command-line flow, BrowserUp load tests are defined in a YAML config file.

  name: Newman
  total_users: 2
    - name: Cart API
      think_time: 10s
      allocation: 50%
      command: newman run postman_collections/cart-api.json
    - name: Refund API
      think_time: 10s
      allocation: 50%
      command: newman run postman_collections/refund-api.json
  stop_after: 15m
    - ramp_to: 100%
      over: 1m
  cluster_type: local

The Important keys:

A scenario represents the definition of work that will happen during a test run.
This top-level key holds an array of profiles. Each profile represents an item that will be
executed during the test.
  • profile
    Each (profile)[profile-settings] is an item that is executed during the test. Each profile must have a name key.
    The name is used to associate metrics captured for this profile with data in our reports.

stages: The stages define a duration, and type of stage for the load test.

images: [optional]
By default, your code is wrapped into a docker container running our custom-base image.
However, if you specify an image, you can use a custom image, based off our image, into which
you have pre-installed any of your own software that will be utilized for the load test.

To specify a custom image, you need to create the image, and ensure it has been uploaded to a registry with builds for each architecture where it will run. This means, at minimum, you need to ensure the docker image is published in the AMD64 architecture which is used for linux cloud machines. For local docker runs, if you use mac, you should also build and push the Arm64 image as well.

Images are referenced by name in BrowserUp, so if you specify a custom image in a profile, it should reference the name in the browserup system.

Creating a test config

When creating a test config for your load test, the first step is to define profile settings for each thing you’d like to run during your test.

Ideally, you’d like to know that the command, when run with the assets in your artict_dir inside our container will 1-Run properly 2-Cause traffic to occur

The BrowserUp Command line utility’s verify command lets you do this.

To verify things are working properly, from within the artifact_dir for your test, at the command line, run a command that you’d like execute succesfully within your container.

For example, if “python3” is the command we’d like to try out in our container, we can run:

browserup verify -v "python3" --artifact-dir=.

With the -v verbose argument, the output of the command will be captured and printed as well as the full HAR file (traffic snapshot) that is captured.

The HAR (HTTP Archive) file is a format used by chrome and other utilities to hold traffic captures. Internally, it is how BrowserUp holds traffic that will be used for metrics.

Without the -v, the verify command will still check for traffic, however it will not print the HAR and output.

The output of the command looks like this:

Once you have succesfully verified your command with your artifact_dir, you can configure a profile settings in the scenario within your test config