Merge pull request #341 from kbst/kgieseking-gke-enhancements

Kgieseking gke enhancements

Author: pst

google/_modules/gke/cluster.tf

@@ -74,15 +74,15 @@ resource "google_container_cluster" "current" {
     }
 
     dynamic "maintenance_exclusion" {
-      for_each = var.maintenance_exclusions
+      for_each = var.maintenance_exclusion_start_time != null ? [1] : []
 
       content {
-        start_time     = maintenance_exclusion.value.start_time
-        end_time       = maintenance_exclusion.value.end_time
-        exclusion_name = maintenance_exclusion.value.exclusion_name
+        start_time     = var.maintenance_exclusion_start_time
+        end_time       = var.maintenance_exclusion_end_time
+        exclusion_name = var.maintenance_exclusion_name
 
         exclusion_options {
-          scope = maintenance_exclusion.value.scope
+          scope = var.maintenance_exclusion_scope
         }
       }
     }

google/_modules/gke/node_pool/main.tf

@@ -14,6 +14,16 @@ resource "google_container_node_pool" "current" {
 
   node_locations = var.node_locations
 
+  dynamic "network_config" {
+    for_each = var.network_config == null ? [] : [1]
+
+    content {
+        enable_private_nodes = var.network_config["enable_private_nodes"]
+        create_pod_range     = var.network_config["create_pod_range"]
+        pod_ipv4_cidr_block  = var.network_config["pod_ipv4_cidr_block"]
+    }
+  }
+
   #
   #
   # Node config
@@ -31,7 +41,7 @@ resource "google_container_node_pool" "current" {
 
     labels = merge(var.labels, var.metadata_labels)
 
-    tags = var.metadata_tags
+    tags = concat(var.metadata_tags, var.instance_tags)
 
     workload_metadata_config {
       mode = var.node_workload_metadata_config

google/_modules/gke/node_pool/variables.tf

@@ -119,6 +119,12 @@ variable "taints" {
   default     = null
 }
 
+variable "instance_tags" {
+  type        = list(string)
+  description = "List of instance tags to apply to nodes."
+  default     = []
+}
+
 variable "node_locations" {
   type        = list(string)
   description = "List of zones in the cluster's region to start worker nodes in. Defaults to cluster's node locations."
@@ -152,3 +158,26 @@ variable "labels" {
   description = "Kubernetes labels to set on the nodes created by the node pool. Merged with Kubestack default labels."
   default     = {}
 }
+
+variable "network_config" {
+  type = object({
+    enable_private_nodes = bool
+    create_pod_range     = bool
+    pod_ipv4_cidr_block  = string
+  })
+  description = "Additional network configuration for the node pool."
+}
+
+variable "ephemeral_storage_local_ssd_config" {
+  type = object({
+    local_ssd_count = number
+  })
+  description = "`ephemeral_storage_local_ssd_config` block, useful for node groups with local SSD. Defaults to `null`"
+  default     = null
+}
+
+variable "labels" {
+  type        = map(string)
+  description = "Kubernetes labels to set on the nodes created by the node pool. Merged with Kubestack default labels."
+  default     = {}
+}

google/_modules/gke/variables.tf

@@ -53,14 +53,24 @@ variable "daily_maintenance_window_start_time" {
   description = "Start time of the daily maintenance window."
 }
 
-variable "maintenance_exclusions" {
-  type = list(object({
-    start_time     = string
-    end_time       = string
-    exclusion_name = string
-    scope          = string
-  }))
-  description = "List of maintenance exclusion configuration to be set on the cluster."
+variable "maintenance_exclusion_start_time" {
+  type = string
+  description = "Maintenance exclusion start time"
+}
+
+variable "maintenance_exclusion_end_time" {
+  type = string
+  description = "Maintenance exclusion end time"
+}
+
+variable "maintenance_exclusion_name" {
+  type = string
+  description = "Maintenance exclusion name"
+}
+
+variable "maintenance_exclusion_scope" {
+  type = string
+  description = "Maintenance exclusion scope"
 }
 
 variable "remove_default_node_pool" {

google/cluster/configuration.tf

@@ -30,7 +30,11 @@ locals {
     "cluster_daily_maintenance_window_start_time",
     "03:00",
   )
-  cluster_maintenance_exclusions = lookup(local.cfg, "cluster_maintenance_exclusions", [])
+
+  cluster_maintenance_exclusion_start_time = lookup(local.cfg, "cluster_maintenance_exclusion_start_time", "")
+  cluster_maintenance_exclusion_end_time = lookup(local.cfg, "cluster_maintenance_exclusion_end_time", "")
+  cluster_maintenance_exclusion_name = lookup(local.cfg, "cluster_maintenance_exclusion_name", "")
+  cluster_maintenance_exclusion_scope = lookup(local.cfg, "cluster_maintenance_exclusion_scope", "")
 
   remove_default_node_pool = lookup(local.cfg, "remove_default_node_pool", true)
 

google/cluster/main.tf

@@ -27,7 +27,11 @@ module "cluster" {
   release_channel    = local.cluster_release_channel
 
   daily_maintenance_window_start_time = local.cluster_daily_maintenance_window_start_time
-  maintenance_exclusions              = local.cluster_maintenance_exclusions
+
+  maintenance_exclusion_start_time = local.cluster_maintenance_exclusion_start_time
+  maintenance_exclusion_end_time = local.cluster_maintenance_exclusion_end_time
+  maintenance_exclusion_name = local.cluster_maintenance_exclusion_name
+  maintenance_exclusion_scope = local.cluster_maintenance_exclusion_scope
 
   remove_default_node_pool = local.remove_default_node_pool
 

google/cluster/node-pool/configuration.tf

@@ -42,5 +42,12 @@ locals {
 
   ephemeral_storage_local_ssd_config = local.cfg["ephemeral_storage_local_ssd_config"]
 
+  guest_accelerator = local.cfg["guest_accelerator"]
+  network_config    = local.cfg["network_config"]
+
+  instance_tags = local.cfg["instance_tags"]
+
+  ephemeral_storage_local_ssd_config = local.cfg["ephemeral_storage_local_ssd_config"]
+
   guest_accelerator = local.cfg["guest_accelerator"]
 }

google/cluster/node-pool/main.tf

@@ -30,7 +30,8 @@ module "node_pool" {
 
   node_workload_metadata_config = local.node_workload_metadata_config
 
-  taints = local.taints
+  taints        = local.taints
+  instance_tags = local.instance_tags
 
   labels = local.labels
 
@@ -40,4 +41,10 @@ module "node_pool" {
   ephemeral_storage_local_ssd_config = local.ephemeral_storage_local_ssd_config
 
   guest_accelerator = local.guest_accelerator
+
+  network_config = local.network_config
+
+  ephemeral_storage_local_ssd_config = local.ephemeral_storage_local_ssd_config
+
+  guest_accelerator = local.guest_accelerator
 }

google/cluster/node-pool/variables.tf

@@ -29,6 +29,8 @@ variable "configuration" {
 
     labels = optional(map(string))
 
+    labels = optional(map(string))
+
     extra_oauth_scopes = optional(list(string))
 
     node_workload_metadata_config = optional(string)
@@ -48,6 +50,28 @@ variable "configuration" {
         max_shared_clients_per_gpu = optional(number)
       }))
     }))
+
+    network_config = optional(object({
+      enable_private_nodes = bool
+      create_pod_range     = bool
+      pod_ipv4_cidr_block  = string
+    }))
+
+    instance_tags = optional(list(string))
+
+    ephemeral_storage_local_ssd_config = optional(object({
+      local_ssd_count = number
+    }))
+
+    guest_accelerator = optional(object({
+      type               = string
+      count              = number
+      gpu_partition_size = optional(string)
+      gpu_sharing_config = optional(object({
+        gpu_sharing_strategy       = optional(string)
+        max_shared_clients_per_gpu = optional(number)
+      }))
+    }))
   }))
 
   description = "Map with per workspace cluster configuration."