k8s-workloads pipelines in the ops instance depend on the availability of the gitlab chart repository on gitlab.com

During an incident with pages: production#2379 (closed) , that resulted in gitlab helm chart repo being unresponsive, all k8s-workloads pipelines were failing on both .com and ops instances, some examples:

We operate under the assumption that if .com is down, we should be able to deploy from ops, so this dependency should be either resolved or documented.