Consider if KCP should log periodically status

What would you like to be added (User Story)?

As an operator Ii would like to easily triage what happened to a cluster's control plane

Detailed Description

https://github.com/kubernetes-sigs/cluster-api/pull/11693 introduced a func tha can generate key/value pairs describing the overall status of the control plane.

Those k/v pair, are then added to "Machine Create" and "Deleting Machine" log lines, thus providing a sort of history of how KCP evolved over time (and why). This is good.

However, unless users have monitoring system on top of CAPI, as of today by looking at logs it is complex to figure out what happened in between "Machine Create" and "Deleting Machine" oprations, e.g. did etcd had issue in the last two hours?

This issue is about discussing options to fill this gap, e.g

• by logging overall status of the control plane e.g. every every 20m (at steady state, every other reconcile)
• by logging overall status when it changes
• ... (any idea striking a balance between getting more info and avoid noise/log size blow up is more than welcome)

Anything else you would like to add?

No response

Label(s) to be applied

/kind feature /area provider/control-plane-kubeadm