Use unique cluster alias as identifier rather than redundant eksID

[linki]

Using eksID() generates redundant and unreadable EKS cluster names, as well as other resources that are based on it.

For example, a cluster with alias playground-2 will have an EKS resource named aws--123456789012--eu-central-1--kube-1 for which the ARN will become a long, unreadable and redundant:

arn:aws:eks:eu-central-1:123456789012:cluster/aws--123456789012--eu-central-1--kube-1

[Update] Since the merge of https://github.com/zalando-incubator/cluster-lifecycle-manager/pull/818 the ARN will look like this:

arn:aws:eks:eu-central-1:123456789012:cluster/kube-1

This isn't too bad, but it's unclear which cluster "kube-1" is. One has to use the AWS account ID to figure that out. Subsequent clusters in the same account would likely use "kube-2", "kube-3" etc. which isn't really helpful.

With this change the cluster will show in the UI as playground-2 and will always be unique as long as we pick unique cluster aliases. The ARN will become more useful and correct:

arn:aws:eks:eu-central-1:123456789012:cluster/playground-2

• [ ] Let's decommission any incompatible EKS cluster before merging this.