GCP deployment: updating node pools cause cluster strange update

[qdelamea-aneo]

Describe the bug
When updating an existing infrastructure deployed on GCP, the container cluster is updated even when only its pools have been modified. Below is an extract of the Terraform plan showing some of the updated resources.

  # module.gke.module.gke[0].google_container_cluster.primary will be updated in-place
  ~ resource "google_container_cluster" "primary" {
        id                          = "projects/armonik-gcp-13469/locations/us-central1/clusters/airflow-bench"
        name                        = "airflow-bench"
        # (30 unchanged attributes hidden)

      ~ addons_config {
          ~ gce_persistent_disk_csi_driver_config {
              ~ enabled = true -> false
            }

            # (9 unchanged blocks hidden)
        }

        # (30 unchanged blocks hidden)
    }

To Reproduce
Steps to reproduce the behavior:

1. Deploy ArmoniK on GCP using the command make.
2. Modify one of the pools and deploy again using make.

[jdongo-aneo]

can you try to set this variable

gce_pd_csi_driver = true