Goal

Move CAPO (cluster-api-provider-openstack) to a stable API.

Why is this important?

Currently OpenShift on OpenStack is using MAPO. This uses objects from the upstream CAPO project under the hood but not the APIs. We would like to start using CAPO and declare MAPO as deprecated and frozen, but before we do that upstream CAPO's own API needs to be declared stable.

Upstream CAPO's API is currently at v1alpha6. There are a number of incompatible changes already planned for the API which have prevented us from declaring it v1beta1. We should make those changes and move towards a stable API.

The changes need to be accompanied by an improvement in test coverage of API versions.

Upstream issues targeted for v1beta1 should be tracked in the v0.7 milestone: https://github.com/kubernetes-sigs/cluster-api-provider-openstack/issues?q=is%3Aopen+is%3Aissue+milestone%3Av0.7

Another option is to switch to cluster-capi-operator if it graduates, which would mean only a single API would be maintained.

Scenarios

N/A. This is purely upstream work for now. We will directly benefit from this work once we switch to CAPO in a future release.

Acceptance Criteria

Upstream CAPO provides a v1beta1 API
Upstream CAPO includes e2e tests using envtest (https://book.kubebuilder.io/reference/envtest.html) which will allow us to avoid breaks in API compatibility

Dependencies (internal and external)

None.

Previous Work (Optional):

N/A