KEP-5667: CRD Map Key Validation

[jpbetz]

We propose adding a x-kubernetes-property-names extension to OpenAPI 3.0.

This extension can be used to validate map keys. For example, to define map
of string to string where the keys and values have a max length of 32:

type: object
properties:
  exampleMap:
    type: object
    x-kubernetes-property-names:
      type: string
      maxLength: 32
    additionalProperties:
      type: string
      maxLength: 32

• Issue link: CRD Map Key Validation #5667

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: jpbetz

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• keps/sig-api-machinery/OWNERS [jpbetz]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[k8s-ci-robot]

@jpbetz: The following tests failed, say /retest to rerun all failed tests or /retest-required to rerun all mandatory failed tests:

Test name	Commit	Details	Required	Rerun command
pull-enhancements-test	bea50d5	link	true	/test pull-enhancements-test
pull-enhancements-verify	bea50d5	link	true	/test pull-enhancements-verify

Full PR test history. Your PR dashboard. Please help us cut down on flakes by linking to an open issue when you hit one in your PR.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here.

[JoelSpeed]

Perhaps a good goal here is to make sure that any changes we build in this direction are compatible with a future move to OpenAPI 3.1?

[jpbetz]

+1. I'll clarify the goals. What we add will be a strict subset of 3.1 capabilities, but put under an extension field to retain OpenAPI 3.0 compatibility. It should be trivially possible to automate a rewrite of a 3.0 schema to a 3.1 schema.

[JoelSpeed]

At the moment it assumes unbounded effectively? So max string length of ~3MiB?

[jpbetz]

Yes, that's right. I'll clarify in the KEP.

[JoelSpeed]

What does this mean? How does defaulting for a key work?

[jpbetz]

TBH I hadn't fully thought this through.

OpenAPI 3.1 makes it possible to declare map keys of different types (not just strings), and to specify defaults, and I wrote this down when I noticed that.

But thinking about it now, I'm struggling to imagine when I'd ever want to default a map key, in practice.

I'll drop this part and just focus on validation support.

[JoelSpeed]

I assume we will need to follow the same pattern for bringing CEL in of introduce the extension in 1.X, but only allow CRDs to write to the extension in 1.X+1 right? To allow for upgrades/rollback cycles

[jpbetz]

Yes, it's the same approach. I'll clarify.