CI fails due to change in behaviour in rancher-webhook

[alexandreLamarre]

Describe the bug

In this commit from a while back:

rancher/webhook@47a27e5#diff-d5ef5135963461ac21aa742cdf581fa0e15e6f760386ca28074a43901301d311R60

new behavior was introduced for handling project admission

if len(values) < 2 {
return nil, fmt.Errorf("unable to retrieve project id from annotation, too few values")
}

This causes our CI definitions that use the old assumptions to fail, see

helm-project-operator/.github/workflows/e2e/scripts/create-project-namespace.sh

Lines 8 to 10 in d17d1d9

	kubectl create namespace e2e-hpo || true
	kubectl label namespace e2e-hpo field.cattle.io/projectId=p-example --overwrite
	sleep "${DEFAULT_SLEEP_TIMEOUT_SECONDS}"

and in prometheus federator https://github.com/rancher/prometheus-federator/blob/1edea93ad2496a8837e4ce78abebac60a14d9e47/.github/workflows/e2e/scripts/create-project-namespace.sh#L9-L11

To Reproduce

Run the script ./.github/workflows/e2e/scripts/create-project-namespace.sh

Expected Result

We need to update the CI/ helm project operator to use this new webhook handling as it is the default in newer kubernetes distributions.