The value for key `cluster-autoscaler.kubernetes.io/scale-down-disabled` placed by MCM is now `true` and not `True`. This typo stopped MCM from disabling CA from scaling down during rolling update.
MCM now marks 1 machine per machineDeployment as Failed at a time in case of healthTimeout. This is introduced to deal with meltdown scenario
machine-controller-manager does now log the Node conditions when it considers Machine as unhealthy (and changes its state to `Unknown`).
Release Notes: