'oc adm release extract --include ...' obsoletes ccoctl's --enable-tech-preview

Description of problem:

ccoctl consumes CredentialsRequests extracted from OpenShift releases and manages secrets associated with those requests for the cluster. Over time, ccoctl has grown a number of CredentialRequest filters, including deletion annotations in CCO-175 and tech-preview annotations in cco#444.

But with OTA-559, 4.14 and later oc adm release extract ... learned about an --included parameter, which allows oc to perform that "will the cluster need this credential?" filtering, and there is no longer a need for ccoctl to perform that filtering, or for ccoctl callers to have to think through "do I need to enable tech-preview CRs for this cluster or not?".

Version-Release number of selected component (if applicable):

4.14 and later.

How reproducible:

100%.

Steps to Reproduce:

$ cat <<EOF >install-config.yaml 
> apiVersion: v1
> platform:
>   gcp:
>     dummy: data
> featureSet: TechPreviewNoUpgrade
> EOF
$ oc adm release extract --included --credentials-requests --install-config install-config.yaml --to credentials-requests quay.io/openshift-release-dev/ocp-release:4.14.0-rc.2-x86_64
$ ccoctl gcp create-all --dry-run --name=test --region=test --project=test --credentials-requests-dir=credentials-requests

Actual results:

ccoctl doesn't dry-run create the TechPreviewNoUpgrade openshift-cluster-api-gcp CredentialsRequest unless you pass it {--enable-tech-preview}}.

Expected results:

ccoctl does dry-run create the TechPreviewNoUpgrade openshift-cluster-api-gcp CredentialsRequest unless you pass it --enable-tech-preview=false.

Additional info:

Longer-term, we likely want to go through some phases of deprecating and maybe eventually removing --enable-tech-preview and the ccoctl-side filtering. But for now, I think we want to pivot to defaulting to true, so that anyone with existing flows that do not include the new --included extraction has an easy way to keep their workflow going (they can set --enable-tech-preview=false). And I think we should backport that to 4.14's ccoctl to simplify OSDOCS-4158's docs#62148. But we're close enough to 4.14's expected GA, that it's worth some consensus-building and alternative consideration, before trying to rush changes back to 4.14 branches.