Create 'pre-existing-gke-cluster' module

[sharabiani]

This PR introduces the pre-existing-gke-cluster module, allowing blueprints to leverage existing Google Kubernetes Engine (GKE) clusters.

Key Changes:

New Module:

Purpose: Facilitates the usage of a pre-existing GKE cluster in HPC Toolkit blueprints.
Location: community/modules/scheduler/pre-existing-gke-cluster
Inputs:

• cluster_name (required): Name of the existing GKE cluster.
• project_id (required): Project that hosts the existing cluster.
• region (required): Region in which to search for the cluster

Manual testing:

1. Created a new GKE cluster.
2. Deployed the provided sample blueprint (hpc-existing-gke).
3. Verified the successful creation of a new node pool in the GKE cluster through Pantheon.

Sample Blueprint:

blueprint_name: hpc-existing-gke

vars:

 project_id: ## Set GCP Project ID Here ##

 deployment_name: existing-cluster-01

 region: us-central1



deployment_groups:

- group: primary

 modules:

 - id: existing_cluster

  source: community/modules/scheduler/pre-existing-gke-cluster

  settings:

   cluster_name: "cluster-01"



 - id: compute_pool

  source: community/modules/compute/gke-node-pool

  use: [existing_cluster]



 - id: job-template

  source: community/modules/compute/gke-job-template

  use: [compute_pool]

  settings:

   image: busybox

   command:

   - echo

   - Hello World

   node_count: 3

  outputs: [instructions]

Additional Notes:

• Made changes to relevant documentation.

[nick-stroud]

You can ignore PR-test-ml-gke failure, this is a new integration test that is just not defined in your pr.

[nick-stroud]

/gcbrun