Contributions and the Contribution Radar

More Tutorials
Contributions and the Contribution Radar

Contributions are a feature from the Open Source Chaos Toolkit. The Contributions Radar allows to easily check what your experiments contributes to.

What are contributions?

From the Chaos Toolkit reference:

Contributions describe the valuable system properties an experiment targets as well as how much they contributes to it. Those properties usually refer to aspects stakeholders care about. Aggregated they offer a powerful metric about the effort and focus on building confidence across the system.

Contributions are defined in your experiment JSON file.

"contributions": {  
    "reliability": "high",
    "security": "none",
    "scalability": "medium"
}

The list of contributions is free, and each value represents the weight of a given contribution. Values must be one of "high", "medium", "low" or "none".

Writing contributions objects

From the previous example, you can see that the experiment explicitely does not address security. Instead of not mentioning security in our contributions object, it has been given a weight of "none".

This allows for two things:

  • Making it clear that the author of the experiment has evaluated all the impacts of the experiment on security, and decided it had none.
  • Making the contributions list for all experiments consistent, organization-wide.

The Contributions Radar

The Contribution Radar is a graphic representation of an experiment's contribution object. It allows for a quick and easy visualisation of the system properties the experiment targets.

In order to make it a valuable tool for quickly comparing the impacts of different experiments, it is highly recommended to make your contributions list consistent for all your experiments.

Let's compare two situations. Here are the Contributions Radars for two different experiments.

Two Contributions Radar that look visually identical, but the legend tells us they represent different things

At first glance, these two radars look identical. But when checking the contributions they represent, you will notice they have nothing in common.

Let's look at two other examples.

Two Contributions Radar that look different, on a chart with the identical legends

These two radars represent the same two experiments as in our previous example. But they are visually very different. That happens because they use the same contributions list, in the exact same order, with a weight of "none" when needed. It allows users to quickly visualize the goals and impacts of an experiment, making the Contributions Radar much more useful.