Skip to main content

Witboost Computational Governance Platform Team Glossary

This glossary page is meant to explain advanced terms used by the Platform Team. If you are a Witboost Computational Governance user or you want to understand the general concepts of the Witboost Computational Governance, we recommend checking out the Computational Governance User Glossary instead.

note

If you're part of the Platform Team, we still encourage going first through the User Glossary as there are concepts a Platform Team member must be familiar as well.

Policy/Metric types​

Policies and metrics can have different types based on the desired behaviour. Here we list the different types of policies with a generic definition, for more information check this section of the documentation:

Trigger types​

  • Active policies: Policies executed as a part of the deployment process of the Control Plane
  • Passive policies: Policies executed independently from the deployment process, either via a schedule or push-based where an external service pushes results to the Computational Governance

Preprocessing​

  • Previous Vs Current policies compare the descriptor of the last deployed version of the resource on the Control Plane against the descriptor of the current version that is going under validation

Engine type​

The Engine on the Witboost Computational Governanace refers to the system or tool used to perform the compliance checks for policies and metrics. By default, Witboost Computational Governance comes with a default Engine using the data validation language CUE. Alongside it, is possible to define remote custom engines that can perform more specialized controls, or engines that harness the power of Large Language Models (LLMs) to declare policies and metrics using natural language.

  • CUE policies: Policies exectued natively by Witboost using a CUE script
  • Remote policies: Policies executed not by the Computational Governance, but by an external service compliant to the Remote Engine API specification and that is called by the Computational Governance.
  • LLM policies: Policies defined using natural languange and executed by contacting a Large Language Model.

Perimeter Resolver​

The Perimeter Resolver is an extension point that retrieves all resources for a specific resource type, and as such it defines which resources are governed by the Computational Governance. Witboost Control Plane provides by default the Perimeter Resolvers for retrieving the Builder project descriptors.