How exactly does a 'Load Balancer' get provisioned on a cloud provider like AWS / GCP when creating a Kubernetes service with a type of LoadBalancer?
Which component is actually responsible for interacting with the cloud provider's API? Or is this something that occurs internally on the cloud provider?
Earlier it was kube-controller-manager but now it is https://kubernetes.io/docs/tasks/administer-cluster/running-cloud-controller/
The cloud controller manager has been taken out of the kube-controller-manager post v1.6.
As of 1.8, it has a service controller which is responsible for loadbalancers on the cloud against services of type LoadBalancer.