Build upstream CAPI

[richardcase]

Build our fork of upstream CAPI using GitHub actions and publish to prime registry (along with slsa provenance).

A couple of things to consider:

• We'll need to do this for some other providers, so let's try and use reusable workflows as much as possible.
• If we can automatically detect a new upstream release and start the rebase/build of the form that would be excellent

[richardcase]

Notes from backlog grooming 2024-01-8.

From CAPI Operator:

• Image override
• Pull location (better UX than specifying fetch URL)
• Air-gapped bundling already done

We need to document the restrictions on when we would bump deps.

[Danil-Grigorev]

CAPI Operator will be able to support image overrides with kubernetes-sigs/cluster-api-operator#349

[richardcase]

As we are waiting EIO to add the prime details lets start publishing to GHCR and we can create a separate issue to enable the prime publication as a follow-up.

[richardcase]

We can make the GHCR images public.

[richardcase]

EIO ticket for creds: https://github.com/rancherlabs/eio/issues/2027

[richardcase]

From planning, just publish to prime registry.