How does creating, developing on and operating K8s cluster differ -
when using Gardener.
OR
directly creating it in GCP / AWS?
I provisioned K8s cluster via Gardner. I now see compute instance created by Gardener (no GKE cluster).
Is that the sole difference, at least in the context of a K8s cluster being provisioned?
How does creating, developing on and operating K8s cluster differ -
when using Gardener.
OR
directly creating it in GCP / AWS?
They are just different ways of managing your cluster/clusters, in essence, they accomplish something very similar with some differences here an there.
Gardener: You have a 'garden' K8s to manage cluster on other cloud providers such as AWS, GCP, Azure, OpenStack. (Running on their server infrastructures like EC2 or GCE)
GCP / AWS. Here you can create your cluster with other tools like kops or kubeadm. The cluster is created in underlying servers in GCE or EC2 (with autoscaling groups, etc)
GKE / EKS / AKS. These are managed services specific to each of the cloud providers like Google, Amazon, and Microsoft. If you create a cluster with one of these you will only be able to manage it through the cloud providers specific API/Services. Their underlying infrastructure also happens to be EC2, GCE, Azure, depending on the cloud provider)
One advantage of Gardener appears to be able to create and manage clusters across multiple clouds from a single point.