Make use of API server dry-run in K8s v1.13

[jessesuen]

Today, we do a kubectl apply --dry-run of all resources, before performing the actual apply, in hopes of catching any manifest errors before deploying anything to the cluster. This is far from foolproof and the dry-run does not catch a lot of errors that only happen during the apply phase, which results in a partial successful deployment. An examples of these may be where a deployment's label selectors do not match the pod template labels.

There is a new API server feature (will be beta in v1.13) which will perform the validation server side, which will improve the experience by performing deeper checks on the user's manifests:

kubernetes/enhancements#576

To support this, we simply need to check if the cluster is running v1.13 and use the new --server-dry-run flag in place of --dry-run.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[leoluz]

Closing this in favor of #11574