Agent waiting for ok to reboot will hammer API server without backoff if API server responds with errors

[invidian]

https://github.com/flatcar-linux/flatcar-linux-update-operator/blob/c180704dde101d2c8313c7beabfdfec24fbe8920/pkg/agent/agent.go#L192-L202 this loop should have some kind of backoff on error, as it may fail in the following conditions:

• getting Node object fails
• creating watcher fails
• watching times out (this happens regularly every 24h #23)

[invidian]

Not only this, but once this happens it also won't stop gracefully after closing the control channel.