Change instance group to regional one

Author: jakubno

packages/cluster/client/main.tf

@@ -20,7 +20,7 @@ resource "google_compute_autoscaler" "default" {
 
   name   = "${var.cluster_name}-autoscaler"
   zone   = var.gcp_zone
-  target = google_compute_instance_group_manager.client_cluster.id
+  target = google_compute_region_instance_group_manager.client_cluster.id
 
   autoscaling_policy {
     max_replicas    = var.cluster_size + var.cluster_auto_scaling_max
@@ -34,8 +34,9 @@ resource "google_compute_autoscaler" "default" {
   }
 }
 
-resource "google_compute_instance_group_manager" "client_cluster" {
-  name = "${var.cluster_name}-ig"
+resource "google_compute_region_instance_group_manager" "client_cluster" {
+  name   = "${var.cluster_name}-ig"
+  region = var.gcp_region
 
   version {
     name              = google_compute_instance_template.client.id

packages/cluster/client/outputs.tf

@@ -7,17 +7,17 @@ output "cluster_tag_name" {
 }
 
 output "instance_group_id" {
-  value = google_compute_instance_group_manager.client_cluster.id
+  value = google_compute_region_instance_group_manager.client_cluster.id
 }
 
 output "instance_group_url" {
-  value = google_compute_instance_group_manager.client_cluster.self_link
+  value = google_compute_region_instance_group_manager.client_cluster.self_link
 }
 
 output "instance_group_name" {
-  value = google_compute_instance_group_manager.client_cluster.name
+  value = google_compute_region_instance_group_manager.client_cluster.name
 }
 
 output "instance_group" {
-  value = google_compute_instance_group_manager.client_cluster.instance_group
+  value = google_compute_region_instance_group_manager.client_cluster.instance_group
 }

packages/cluster/client/variables.tf

@@ -8,6 +8,11 @@ variable "environment" {
   type        = string
 }
 
+variable "gcp_region" {
+  description = "The GCP region in which the server cluster will be created (e.g. us-central1-a)."
+  type        = string
+}
+
 variable "gcp_zone" {
   description = "The GCP zone in which the server cluster will be created (e.g. us-central1-a)."
   type        = string

packages/cluster/main.tf

@@ -136,7 +136,9 @@ module "client_cluster" {
   cluster_auto_scaling_max = var.client_cluster_auto_scaling_max
   cluster_size             = var.client_cluster_size
   cluster_tag_name         = var.cluster_tag_name
-  gcp_zone                 = var.gcp_zone
+
+  gcp_region = var.gcp_region
+  gcp_zone   = var.gcp_zone
 
   machine_type = var.client_machine_type
   image_family = var.client_image_family

packages/nomad/api.hcl

@@ -1,5 +1,4 @@
 job "api" {
-  datacenters = ["${gcp_zone}"]
   node_pool = "api"
   priority = 90
 

packages/nomad/clickhouse.hcl

@@ -1,5 +1,4 @@
 job "clickhouse" {
-  datacenters = ["${zone}"]
   type        = "service"
   node_pool = "api"
 

packages/nomad/client-proxy.hcl

@@ -1,5 +1,4 @@
 job "client-proxy" {
-  datacenters = ["${gcp_zone}"]
   node_pool = "api"
 
   priority = 80

packages/nomad/docker-reverse-proxy.hcl

@@ -1,8 +1,3 @@
-variable "gcp_zone" {
-  type    = string
-  default = ""
-}
-
 variable "gcp_region" {
   type    = string
   default = ""
@@ -54,7 +49,6 @@ variable "google_service_account_secret" {
 }
 
 job "docker-reverse-proxy" {
-  datacenters = [var.gcp_zone]
   node_pool = "build"
 
 

packages/nomad/logs-collector.hcl

@@ -1,5 +1,4 @@
 job "logs-collector" {
-  datacenters = ["${gcp_zone}"]
   type        = "system"
   node_pool    = "all"
 

packages/nomad/loki.hcl

@@ -1,5 +1,4 @@
 job "loki" {
-  datacenters = ["${gcp_zone}"]
   type        = "service"
   node_pool = "api"
 

packages/nomad/main.tf

@@ -37,7 +37,6 @@ resource "nomad_job" "api" {
     otel_collector_grpc_endpoint  = "localhost:4317"
     loki_address                  = "http://loki.service.consul:${var.loki_service_port.port}"
     logs_collector_address        = "http://localhost:${var.logs_proxy_port.port}"
-    gcp_zone                      = var.gcp_zone
     port_name                     = var.api_port.name
     port_number                   = var.api_port.port
     api_docker_image              = var.api_docker_image_digest
@@ -62,7 +61,6 @@ resource "nomad_job" "api" {
 resource "nomad_job" "redis" {
   jobspec = templatefile("${path.module}/redis.hcl",
     {
-      gcp_zone    = var.gcp_zone
       port_number = var.redis_port.port
       port_name   = var.redis_port.name
     }
@@ -74,7 +72,6 @@ resource "nomad_job" "docker_reverse_proxy" {
 
   hcl2 {
     vars = {
-      gcp_zone                      = var.gcp_zone
       image_name                    = var.docker_reverse_proxy_docker_image_digest
       postgres_connection_string    = data.google_secret_manager_secret_version.postgres_connection_string.secret_data
       google_service_account_secret = var.docker_reverse_proxy_service_account_key
@@ -94,7 +91,6 @@ resource "nomad_job" "client_proxy" {
     {
       update_stanza = var.api_machine_count > 1
 
-      gcp_zone           = var.gcp_zone
       port_name          = var.client_proxy_port.name
       port_number        = var.client_proxy_port.port
       health_port_number = var.client_proxy_health_port.port
@@ -112,7 +108,6 @@ resource "nomad_job" "session_proxy" {
 
   hcl2 {
     vars = {
-      gcp_zone                   = var.gcp_zone
       session_proxy_port_number  = var.session_proxy_port.port
       session_proxy_port_name    = var.session_proxy_port.name
       session_proxy_service_name = var.session_proxy_service_name
@@ -206,8 +201,6 @@ resource "nomad_job" "otel_collector" {
     grafana_otlp_url             = data.google_secret_manager_secret_version.grafana_otlp_url.secret_data
     grafana_username             = data.google_secret_manager_secret_version.grafana_username.secret_data
     consul_token                 = var.consul_acl_token_secret
-
-    gcp_zone = var.gcp_zone
   })
 }
 
@@ -286,8 +279,6 @@ data "google_secret_manager_secret_version" "grafana_logs_collector_api_token" {
 
 resource "nomad_job" "logs_collector" {
   jobspec = templatefile("${path.module}/logs-collector.hcl", {
-    gcp_zone = var.gcp_zone
-
     logs_port_number        = var.logs_proxy_port.port
     logs_health_port_number = var.logs_health_proxy_port.port
     logs_health_path        = var.logs_health_proxy_port.health_path
@@ -306,12 +297,6 @@ data "google_storage_bucket_object" "orchestrator" {
   bucket = var.fc_env_pipeline_bucket_name
 }
 
-
-data "google_compute_machine_types" "client" {
-  zone   = var.gcp_zone
-  filter = "name = \"${var.client_machine_type}\""
-}
-
 data "external" "orchestrator_checksum" {
   program = ["bash", "${path.module}/checksum.sh"]
 
@@ -322,7 +307,6 @@ data "external" "orchestrator_checksum" {
 
 resource "nomad_job" "orchestrator" {
   jobspec = templatefile("${path.module}/orchestrator.hcl", {
-    gcp_zone         = var.gcp_zone
     port             = var.orchestrator_port
     proxy_port       = var.orchestrator_proxy_port
     environment      = var.environment
@@ -362,7 +346,6 @@ resource "nomad_job" "template_manager" {
 
     gcp_project = var.gcp_project_id
     gcp_region  = var.gcp_region
-    gcp_zone    = var.gcp_zone
     port        = var.template_manager_port
     environment = var.environment
 
@@ -377,10 +360,9 @@ resource "nomad_job" "template_manager" {
     logs_collector_address       = "http://localhost:${var.logs_proxy_port.port}"
   })
 }
+
 resource "nomad_job" "loki" {
   jobspec = templatefile("${path.module}/loki.hcl", {
-    gcp_zone = var.gcp_zone
-
     // We use colocation 2 here to ensure that there are at least 2 nodes for API to do rolling updates.
     // It might be possible there could be problems if we are rolling updates for both API and Loki at the same time., so maybe increasing this to > 3 makes sense.
     prevent_colocation = var.api_machine_count > 2
@@ -418,7 +400,6 @@ resource "google_storage_hmac_key" "clickhouse_hmac_key" {
 # Add this with your other Nomad jobs
 resource "nomad_job" "clickhouse" {
   jobspec = templatefile("${path.module}/clickhouse.hcl", {
-    zone                = var.gcp_zone
     clickhouse_version  = "25.1.5.31" # Or make this a variable
     gcs_bucket          = google_storage_bucket.clickhouse_bucket.name
     gcs_folder          = "clickhouse-data"

packages/nomad/orchestrator.hcl

@@ -1,6 +1,5 @@
 job "orchestrator" {
   type = "system"
-  datacenters = ["${gcp_zone}"]
 
   priority = 90
 

packages/nomad/otel-collector.hcl

@@ -1,5 +1,4 @@
 job "otel-collector" {
-  datacenters = ["${gcp_zone}"]
   type        = "system"
   node_pool   = "all"
 

packages/nomad/redis.hcl

@@ -1,5 +1,4 @@
 job "redis" {
-  datacenters = ["${gcp_zone}"]
   node_pool = "api"
   type = "service"
   priority = 95

packages/nomad/session-proxy.hcl

@@ -1,7 +1,3 @@
-variable "gcp_zone" {
-  type = string
-}
-
 variable "session_proxy_port_number" {
   type = number
 }
@@ -24,7 +20,6 @@ variable "nginx_conf" {
 
 job "session-proxy" {
   type = "system"
-  datacenters = [var.gcp_zone]
 
   priority = 80
 

packages/nomad/template-manager.hcl

@@ -1,5 +1,4 @@
 job "template-manager" {
-  datacenters = ["${gcp_zone}"]
   node_pool  = "build"
   priority = 70