Implement read current state of a managed topology

[fabriziopandini]

Detailed Description
This is part of the activities for the implementation of the ClusterClass proposal.

As part of the ManageTopologyController reconcile loop, we want to read a ClusterClass and all the referenced templates

This step should return a struct including

• Cluster (from scaffolding)
• ClusterInfrastructure (Get*)
• ClusterControlPlane (Get*)
• ControlPlaneMachineInfrastructureTemplate (Get*)
• List of machine deployments with the cluster.x-k8s.io/cluster-name and cluster.x-k8s.io/topology label
  • Then, for each MachineDeployment
    • MachineBootstrapTemplate (Get*)
    • MachineInfrastructureTemplate (Get*)

Anything else you would like to add:

This could be implemented only after #4933 is completed.

Notes:

• Before using a reference it is important to call ConvertReferenceAPIContract in order to properly handle APIVersions/contract - -- upgrades.
• TBD if in case a reference is converted to a newer API contract, the Cluster/MachineDeployment/ControlPlane objects containing the reference should be patched (the existing controllers are already taking care of this; in case we are going to implement this, feel free to tackle this in a separated PR).
• ControlPlaneTemplate object might or might not have a MachineInfrastructureTemplate; in order to discover this, it is required to inspect the template spec using unstructured utils (See 🌱 Extend the ControlPlane contract defining version and infrastructure machine fields for ClusterClass compat #4949 for the related contract changes).
• All the get reading templates/provider types (Get*) must use the unstructured caching client and return unstructured; please note that the external package already provides a Get utility that can be used for this scope.
• After reading the ClusterInfrastructure object and the ClusterControlPlane object it is required to set up a watch on the corresponding types.
• In case of missing objects the read operation should fail; this probably will be changed in future if/when implementing support for reconciling a subset of a topology.
• Other errors (e.g. API server connection errors) must always trigger failure for the operation.

/kind feature
/milestone v0.4

[fabriziopandini]

/assign @killianmuldoon

[k8s-ci-robot]

@fabriziopandini: GitHub didn't allow me to assign the following users: killianmuldoon.

Note that only kubernetes-sigs members, repo collaborators and people who have commented on this issue/PR can be assigned. Additionally, issues/PRs can only have 10 assignees at the same time.
For more information please see the contributor guide

In response to this:

/assign @killianmuldoon

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/test-infra repository.

[fabriziopandini]

(assigning to myself given that Killian is not yet kubernetes-sigs member)

[sbueringer]

@fabriziopandini Just fyi, afaik you cannot assign killian but @killianmuldoon can assign himself. Just found that out with someone from my old company.

Apparently you cannot assign somebody outside of the org, except:

• he does it himself (Example: v1alpha4: Update Readme cluster-api-provider-openstack#943)
• he wrote something in the issue before