feat: add support for gke node group auto provisioning (#14)

* feat: add support for gke node group auto provisioning * fix: remove defaults * fix: remove defaults * doc: update docs
ouzi-dev · Mar 27, 2020 · 8c78bd0 · 8c78bd0
1 parent dd13b3c
commit 8c78bd0
Show file tree

Hide file tree

Showing 8 changed files with 159 additions and 9 deletions.
diff --git a/.gitignore b/.gitignore
@@ -1,4 +1,4 @@
 bootstrap/.terraform
-.terraform
+.terraform/
 terraform.tfstate.backup
 node_modules/
diff --git a/README.md b/README.md
@@ -1,18 +1,84 @@
 # gke-terraform
 
-GKE Terraform module
+A GKE Terraform module that creates a GKE cluster with various options.
+
+## Features
+
+* zonal or regional 
+* master, node and service CIDR range
+* node Scopes
+* authorized cidr blocks for the API
+* kubernetes version 
+* GKE features can be turned on or off 
+  * HPA
+  * LB
+  * Dashboard
+  * Network Policies
+  * Calico
+  * Logging Service
+  * Monitoring Service
+  * Istio
+  * NodeGroup Auto Provisioning - allows GKE to auto provision Node Groups and autoscale them without having to specify node groups at all 
+* optional NodeGroups
+  * support autoscaling within the node group
 
 ## How to Use
 
+### With Cluster NodeGroup AutoProvisioning
+
 ```
+data "google_compute_zones" "available" {}
+
 ## Modules
 module "gke-cluster" {
-  source  = "git@github.com:ouzi-dev/gke-terraform.git?ref=v0.3"
+  source = "github.com/ouzi-dev/gke-terraform.git?ref=v0.9.0"
   region  = var.gcloud_region
   project = var.gcloud_project
 
-  cluster_name       = local.gke_name
-  zones              = var.gke_zones
+  cluster_name = var.gke_name
+  zones = slice(data.google_compute_zones.available.names, 0, var.gke_num_of_zones)
+
+  node_cidr_range    = var.gke_node_cidr_range
+  pod_cidr_range     = var.gke_pod_cidr_range
+  service_cidr_range = var.gke_service_cidr_range
+  master_cidr_range  = var.gke_master_cidr_range
+  gke_node_scopes    = var.gke_node_scopes
+  auth_cidr_blocks   = var.gke_auth_cidr_blocks
+  kubernetes_version = var.gke_kubernetes_version
+
+  cluster_autoscaling = var.cluster_autoscaling
+  cluster_autoscaling_min_cpu = var.cluster_autoscaling_min_cpu
+  cluster_autoscaling_max_cpu = var.cluster_autoscaling_max_cpu
+  cluster_autoscaling_min_memory = var.cluster_autoscaling_min_memory
+  cluster_autoscaling_max_memory = var.cluster_autoscaling_max_memory
+
+  daily_maintenance                   = var.gke_daily_maintenance
+  disable_hpa                         = var.gke_disable_hpa
+  disable_lb                          = var.gke_disable_lb
+  disable_dashboard                   = var.gke_disable_dashboard
+  disable_network_policy              = var.gke_disable_network_policy
+  disable_istio                       = var.gke_disable_istio
+  enable_calico                       = var.gke_enable_calico
+  authenticator_groups_security_group = var.gke_authenticator_groups_security_group
+  init_nodes                          = var.gke_init_nodes
+
+  logging_service    = var.logging_service
+  monitoring_service = var.monitoring_service
+}
+``` 
+### Without Cluster NodeGroup AutoProvisioning
+
+```
+data "google_compute_zones" "available" {}
+
+module "gke-cluster" {
+  source = "github.com/ouzi-dev/gke-terraform.git?ref=v0.9.0"
+  region  = var.gcloud_region
+  project = var.gcloud_project
+
+  cluster_name = var.gke_name
+  zones = slice(data.google_compute_zones.available.names, 0, var.gke_num_of_zones)
+
   node_cidr_range    = var.gke_node_cidr_range
   pod_cidr_range     = var.gke_pod_cidr_range
   service_cidr_range = var.gke_service_cidr_range
@@ -26,14 +92,20 @@ module "gke-cluster" {
   machine_is_preemptible = var.gke_machine_is_preemptible
   min_nodes              = var.gke_min_nodes
   max_nodes              = var.gke_max_nodes
+  max_surge              = var.max_surge
+  max_unavailable        = var.max_unavailable
 
   daily_maintenance                   = var.gke_daily_maintenance
   disable_hpa                         = var.gke_disable_hpa
   disable_lb                          = var.gke_disable_lb
   disable_dashboard                   = var.gke_disable_dashboard
   disable_network_policy              = var.gke_disable_network_policy
+  disable_istio                       = var.gke_disable_istio
   enable_calico                       = var.gke_enable_calico
   authenticator_groups_security_group = var.gke_authenticator_groups_security_group
   init_nodes                          = var.gke_init_nodes
+
+  logging_service    = var.logging_service
+  monitoring_service = var.monitoring_service
 }
-``` 
+```
diff --git a/main.tf b/main.tf
@@ -30,10 +30,20 @@ module "cluster" {
   monitoring_service                  = var.monitoring_service
   disable_istio                       = var.disable_istio
   istio_config_auth                   = var.istio_config_auth
+  cluster_autoscaling                 = var.cluster_autoscaling
+  cluster_autoscaling_profile         = var.cluster_autoscaling_profile
+  cluster_autoscaling_min_cpu         = var.cluster_autoscaling_min_cpu
+  cluster_autoscaling_max_cpu         = var.cluster_autoscaling_max_cpu
+  cluster_autoscaling_min_memory      = var.cluster_autoscaling_min_memory
+  cluster_autoscaling_max_memory      = var.cluster_autoscaling_max_memory
+  cluster_autoscaling_gke_scopes      = var.gke_node_scopes
 }
 
 module "default_workers" {
-  source                 = "./modules/gke-workers"
+  source = "./modules/gke-workers"
+
+  enabled = var.cluster_autoscaling ? false : true
+
   region                 = var.region
   group_name             = "default"
   zones                  = var.zones

diff --git a/modules/gke-cluster/main.tf b/modules/gke-cluster/main.tf
@@ -40,6 +40,24 @@ resource "google_container_cluster" "k8s-cluster" {
   remove_default_node_pool = true
   initial_node_count       = 1
 
+  cluster_autoscaling {
+    enabled             = var.cluster_autoscaling
+    autoscaling_profile = var.cluster_autoscaling_profile
+    resource_limits {
+      resource_type = "cpu"
+      minimum       = var.cluster_autoscaling_min_cpu
+      maximum       = var.cluster_autoscaling_max_cpu
+    }
+    resource_limits {
+      resource_type = "memory"
+      minimum       = var.cluster_autoscaling_min_memory
+      maximum       = var.cluster_autoscaling_max_memory
+    }
+    auto_provisioning_defaults {
+      oauth_scopes = var.cluster_autoscaling_gke_scopes
+    }
+  }
+
   maintenance_policy {
     daily_maintenance_window {
       start_time = var.daily_maintenance

diff --git a/modules/gke-cluster/variables.tf b/modules/gke-cluster/variables.tf
@@ -70,4 +70,21 @@ variable "disable_istio" {
 }
 
 variable "istio_config_auth" {
-}
+}
+
+variable "cluster_autoscaling" {
+  type    = bool
+  default = false
+}
+
+variable "cluster_autoscaling_profile" {}
+
+variable "cluster_autoscaling_min_cpu" {}
+
+variable "cluster_autoscaling_max_cpu" {}
+
+variable "cluster_autoscaling_min_memory" {}
+
+variable "cluster_autoscaling_max_memory" {}
+
+variable "cluster_autoscaling_gke_scopes" {}
diff --git a/modules/gke-workers/main.tf b/modules/gke-workers/main.tf
@@ -1,4 +1,5 @@
 resource "google_container_node_pool" "k8s-worker-pool" {
+  count              = var.enabled ? 1 : 0
   provider           = google-beta
   name               = "${var.gke_cluster_name}-${var.group_name}"
   location           = var.region

diff --git a/modules/gke-workers/variables.tf b/modules/gke-workers/variables.tf
@@ -52,4 +52,9 @@ variable "max_surge" {
 }
 
 variable "max_unavailable" {
+}
+
+variable "enabled" {
+  type    = bool
+  default = true
 }
diff --git a/variables.tf b/variables.tf
@@ -137,4 +137,31 @@ variable "disable_istio" {
 
 variable "istio_config_auth" {
   default = "AUTH_MUTUAL_TLS"
-}
+}
+
+# https://cloud.google.com/kubernetes-engine/docs/how-to/node-auto-provisioning
+variable "cluster_autoscaling" {
+  default = false
+}
+
+# OPTIMIZE_UTILIZATION or BALANCED
+variable "cluster_autoscaling_profile" {
+  type = string
+  default = "OPTIMIZE_UTILIZATION"
+}
+
+variable "cluster_autoscaling_min_cpu" {
+  type = number
+}
+
+variable "cluster_autoscaling_max_cpu" {
+  type = number
+}
+
+variable "cluster_autoscaling_min_memory" {
+  type = number
+}
+
+variable "cluster_autoscaling_max_memory" {
+  type = number
+}