Changes Uses version (v0.1.4) of infoblox module with pinned provider that introduced a bug Used version (v0.1.5) of cluster module that enables gce_pd_csi_driver and changes hub-db-dir storage to standard-rwo classEnabling this makes standard_rwo storage class available, which resolves standard volumes from not binding, possibly due to a timing issue. For ClassStorage standard volumeBindingMode: Immediate for standard_rwo -> volumeBindingMode: WaitForFirstConsumer See more https://stackoverflow.com/questions/72553227/gke-can-t-scale-up-nodes-due-of-persistentvolume Adds kubernetes_version as a variable and tests for 1.24. 1.25 breaks the pod-schedulerAt the moment we are using helm chart 1.20… not sure if this is fixed in 2.0, but this is likely coming from a newer k8s version and a depreciation of v1beta1 policy/v1beta1 PodDisruptionBudget is deprecated in v1.21+, unavailable in v1.25+; use policy/v1 PodDisruptionBudget See Discourse thread