Resource Definitions

Resource definitions describe the schema for the different resource kinds available in a Cup instance. When you interact with the Cup API, you do so through the various configured resource definitions.

A Cup resource is comprised of a few components to build the entire definition. These components and concepts are taken directly from Kubernetes and its custom resource definitions (CRD). Each definition consists of:

• A group
• A set of kind names
• Multiple versioned JSON schema definitions
• Some extra user defined arbitrary metadata

Groups

Groups are a top-level construct for grouping related kinds of resource definitions together. Usually a group name identifies an organization or domain for a set of related resource types.

For example, flipt.io is the group for Flipt feature flag related resources.

Kind names

Kinds are the the unique name identifier for a resource scoped within a particular group. As with Kubernetes CRDs, Cup's resource definitions required you to declare a few variations on the kinds name.

These variations are:

• Kind (canonical kind name in Title-case)
• singular (lowercase term referring to a single instance of the kind)
• plural (lowercase term referring to two or more instances of the kind)

For example, for a Flipt feature flag we have the names:

• Kind is Flag
• singular is flag
• plural is flags

These different forms can be used by downstream tools to provide meaningful and readable interfaces when interacting with these resource kinds.

Versioned JSON Schema

At their core, the resource definitions are comprised of schemas for the fields the resource kinds contain. Cup uses JSON Schema (opens in a new tab) as the schema definition language.

Cup organizes schemas into separate versions. This is important, because over time resources can and will change. Versions allow definition authors to provide guarantees to downstream consumers regarding the shape of resources.