Context:

Problem: It's a strong UX requirement having clear failure scenarios pointing to root problem captured in the API status of the scalable resource without having to look at all Machines.

CAPI doesn't bubble up individual machine failures into MachineDeployments https://github.com/kubernetes-sigs/cluster-api/issues/3486

https://github.com/kubernetes-sigs/cluster-api/issues/5635

We workaround this in hypershift via https://github.com/openshift/hypershift/pull/1907

Previous work

https://github.com/kubernetes-sigs/cluster-api/pull/6218

https://github.com/kubernetes-sigs/cluster-api/pull/6025

DoD:

Ensure and implement an upstream solution to solve the "Problem" described above and let downstream rely on it.