Cluster-api: Add conditions for MachinePool object

[CecileRobertMichon]

Implement https://github.com/kubernetes-sigs/cluster-api/blob/master/docs/proposals/20200506-conditions.md for the MachinePool resource.

Related: #3476

/kind feature

[CecileRobertMichon]

/help

[k8s-ci-robot]

@CecileRobertMichon:
This request has been marked as needing help from a contributor.

Please ensure the request meets the requirements listed here.

If this request no longer meets these requirements, the label can be removed
by commenting with the /remove-help command.

In response to this:

/help

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.

[srm09]

/assign @srm09

[srm09]

As part of this issue, I was thinking of adding the following conditions:

• Bootstrap Ready condition which reports the state of the Bootstrap config object
• Infrastructure Ready condition which reports the state of the Infra config object
• A new condition which reports the state of the replicas/machines associated to the MachinePool object.

@CecileRobertMichon do these make sense?

cc: @fabriziopandini

[fabriziopandini]

Machine pool is an interesting abstraction because from 1 side it has a single infrastructure/bootstrap, from the other side it manages multiple machines.

• I'm +1 in having BootstrapReady, InfrastructureReady
• I think we should have MachinesSpecUpToDate & Resized (similar to KCP) to provide an overview of the status of the underlying machines
• I would be interesting to get a MachineReady condition to provide an overview of the operational state of machines, but I'm not sure if/how this can be achieved

[srm09]

The machinepool controller has the following logic to check if the nodes are available in the reconcileNodeRefs loop.
https://github.com/kubernetes-sigs/cluster-api/blob/73624ed082e948e693a14484549b3db53c1b7ae1/exp/controllers/machinepool_controller_noderef.go#L180-L182

This seems like a place to populate condition data for providing an overview of the operational state of the underlying machines involved.

[CecileRobertMichon]

Sounds good to me. MachinePool should be pretty similar to a Machine in that it's a single infrastructure resource but we can add a new one to track the state of the instances/replicas/nodes. I think we should avoid calling them "Machines" to avoid confusion with Machines in Machine deployments.