Following the rotation of the API key, new autoscaled nodes were affected by the taint.

node.cloudprovider.kubernetes.io/uninitialized=true:NoSchedule

This occurred because the Hetzner Cloud Controller Manager was unable to authenticate with the Hetzner API due to an invalid API token.
I needed to restart the following deployments to ensure they would receive the new token:
• hcloud-cloud-controller-manager
• cluster-autoscaler

After the restart, new nodes initialized successfully without the taint.

Would it be possible to add a new switch “—restart” or similar to restart the deployments so that they would receive the new token if the API key is changed?
Adding this logic to the code would be time-consuming, as it would require constant monitoring of the API key change, which is a rare occurrence.