Autoscaling Node Deployment is scaled up upon every tf apply

[jdftapi]

Provider Version: 5.3.3

Observed Behaviour

We have created a node deployment like this:

resource "metakube_node_deployment" "storage-backup-nodes" {
  name = "storage-backup-nodes"
  cluster_id = metakube_cluster.cluster.id
  spec {
    min_replicas = 0
    max_replicas = 3
    template {
      cloud {
        openstack {
          flavor = var.storage_backup_nodepool.flavor
          image = var.storage_backup_nodepool.image
          use_floating_ip = false
        }
      }
      operating_system {
        ubuntu {}
      }
      versions {
        kubelet = 1.28.11
      }
      labels = {
        "labels.ftapi.com/run" = "storage-backup"
        "labels.ftapi.com/disktype" = "network-shared"
      }
      taints {
        key = "run"
        value = "storage-backup"
        effect = "NoSchedule"
      }
    }
  }
}

It is autoscaling from 0 to 3. Whenever we run terraform apply, terraform wants to change the replicas from the current value (0 for most of the time) to the max_replicas value (3). We explicitly have NOT set a value for replicas.

  # module.metakube_prod.module.cluster_prod-fes1-1.metakube_node_deployment.storage-backup-nodes will be updated in-place
  ~ resource "metakube_node_deployment" "storage-backup-nodes" {
        id                 = "storage-backup-nodes"
        name               = "storage-backup-nodes"
        # (4 unchanged attributes hidden)

      ~ spec {
          ~ replicas     = 0 -> 3
            # (2 unchanged attributes hidden)

            # (1 unchanged block hidden)
        }
    }

Expected Behaviour

Since only min_replicas and max_replicas are defined, we expect terraform to ignore the replicas value.

[phiphi282]

This will be fixed with version v5.4.0 We decided to remove the default entirely and instead make it mandatory to either set replicas or min and max replicas See #108