[DEV] Teach CI how to install Helm charts into a GKE managed cluster

[bcfriesen]

It is non-trivial to evaluate changes to the Helm chart deployment recipe without access to a Kubernetes environment. One way to address this is to provide a cloud-based Kubernetes-as-a-service resource to the OpenCHAMI project. Such a resource will also enable automated integration testing for the project.

I have configured a GKE Autopilot cluster for this purpose. It uses a service account, whose credentials will be stored as a GitHub Secret in this repository, to install Helm charts into this cluster as part of the CI pipeline.

There is nothing special about GKE per se, except that it was the easiest of the K8s-aaS cloud resources available for me to configure. One could accomplish the same task with, e.g., Amazon EKS or Microsoft AKS.