feat(infra): add cost optimization and performance variables for Cloud Run and Redis VPC connector

infra/dcp/main.tf

@@ -97,6 +97,8 @@ locals {
     max_instances                   = var.datacommons_services_max_instances
     cpu                             = var.datacommons_services_cpu
     memory                          = var.datacommons_services_memory
+    cpu_idle                        = var.datacommons_services_cpu_idle
+    startup_cpu_boost               = var.datacommons_services_startup_cpu_boost
     google_analytics_tag            = var.datacommons_services_google_analytics_tag_id
     enable_mcp                      = var.datacommons_services_enable_mcp
     search_scope                    = var.datacommons_services_mcp_search_scope
@@ -121,8 +123,10 @@ locals {
     location_id             = var.redis_location_id
     alternative_location_id = var.redis_alternative_location_id
     replica_count           = var.redis_replica_count
-    vpc_network_name        = var.redis_vpc_network_name
-    vpc_connector_cidr      = var.redis_vpc_connector_cidr
+    vpc_network_name            = var.redis_vpc_network_name
+    vpc_connector_cidr          = var.redis_vpc_connector_cidr
+    vpc_connector_min_instances = var.redis_vpc_connector_min_instances
+    vpc_connector_max_instances = var.redis_vpc_connector_max_instances
   }
 
   ingestion_config = {
@@ -149,6 +153,8 @@ locals {
     # Workflow & Helper Service
     workflow_lock_acquisition_timeout = var.ingestion_workflow_lock_acquisition_timeout
     helper_service_image              = coalesce(var.ingestion_helper_service_image, "gcr.io/datcom-ci/datacommons-ingestion-helper:${var.dcp_version}")
+    helper_service_cpu_idle           = var.ingestion_helper_service_cpu_idle
+    helper_service_startup_cpu_boost  = var.ingestion_helper_service_startup_cpu_boost
 
     # Dataflow Network & Scaling Configuration
     dataflow_ip_configuration    = var.ingestion_dataflow_ip_configuration

infra/dcp/modules/datacommons_services/main.tf

@@ -40,6 +40,8 @@ resource "google_cloud_run_v2_service" "dc_web_service" {
     containers {
       image = var.image
       resources {
+        cpu_idle          = var.cpu_idle
+        startup_cpu_boost = var.startup_cpu_boost
         limits = {
           cpu    = var.cpu
           memory = var.memory

infra/dcp/modules/datacommons_services/variables.tf

@@ -35,6 +35,18 @@ variable "memory" {
   type = string
 }
 
+variable "cpu_idle" {
+  type        = bool
+  description = "When true, CPU is only allocated during request processing (cheaper for low-traffic services). When false, CPU is always allocated (better performance, avoids cold starts)."
+  default     = false
+}
+
+variable "startup_cpu_boost" {
+  type        = bool
+  description = "Temporarily boost CPU allocation during container startup to reduce cold start latency."
+  default     = true
+}
+
 variable "min_instances" {
   type = number
 }

infra/dcp/modules/ingestion/helper_service/main.tf

@@ -18,6 +18,10 @@ resource "google_cloud_run_v2_service" "ingestion_helper" {
     timeout = "1800s"
     containers {
       image = var.image
+      resources {
+        cpu_idle          = var.cpu_idle
+        startup_cpu_boost = var.startup_cpu_boost
+      }
 
       env {
         name  = "PROJECT_ID"

infra/dcp/modules/ingestion/helper_service/variables.tf

@@ -71,6 +71,20 @@ variable "ingestion_artifacts_path" {
   description = "Path where pre-processed files are placed for the next stage"
 }
 
+variable "cpu_idle" {
+  type        = bool
+  description = "When true, CPU is only allocated during request processing (cheaper for low-traffic services). When false, CPU is always allocated."
+  default     = false
+  nullable    = false
+}
+
+variable "startup_cpu_boost" {
+  type        = bool
+  description = "Temporarily boost CPU allocation during container startup to reduce cold start latency."
+  default     = true
+  nullable    = false
+}
+
 variable "skip_container_restarts" {
   type        = bool
   description = "Set to true to skip updating container restart timestamps, speeding up terraform apply when container images have not changed."

infra/dcp/modules/redis/main.tf

@@ -27,6 +27,6 @@ resource "google_vpc_access_connector" "connector" {
   region        = var.region
   network       = var.vpc_network_id
   ip_cidr_range = var.vpc_connector_cidr
-  min_instances = 2
-  max_instances = 10
+  min_instances = var.vpc_connector_min_instances
+  max_instances = var.vpc_connector_max_instances
 }

infra/dcp/modules/redis/variables.tf

@@ -44,3 +44,15 @@ variable "enable_connector" {
   type    = bool
   default = true
 }
+
+variable "vpc_connector_min_instances" {
+  type        = number
+  description = "Minimum number of VPC Access Connector instances."
+  default     = 2
+}
+
+variable "vpc_connector_max_instances" {
+  type        = number
+  description = "Maximum number of VPC Access Connector instances."
+  default     = 10
+}

infra/dcp/modules/stack/main.tf

@@ -199,6 +199,8 @@ module "ingestion_helper_service" {
   redis_port               = var.redis_config.enable && length(module.redis) > 0 ? tostring(module.redis[0].redis_port) : ""
   ingestion_artifacts_path = "${var.ingestion_config.ingestion_artifacts_path}/metadata"
   skip_container_restarts  = var.global.skip_container_restarts
+  cpu_idle                 = var.ingestion_config.helper_service_cpu_idle
+  startup_cpu_boost        = var.ingestion_config.helper_service_startup_cpu_boost
 }
 
 
@@ -245,8 +247,10 @@ module "redis" {
   alternative_location_id = var.redis_config.alternative_location_id
   replica_count           = var.redis_config.replica_count
   vpc_network_id          = data.google_compute_network.default.id
-  vpc_connector_cidr      = var.redis_config.vpc_connector_cidr
-  enable_connector        = true
+  vpc_connector_cidr          = var.redis_config.vpc_connector_cidr
+  vpc_connector_min_instances = var.redis_config.vpc_connector_min_instances
+  vpc_connector_max_instances = var.redis_config.vpc_connector_max_instances
+  enable_connector            = true
 }
 
 module "auth" {
@@ -270,6 +274,8 @@ module "datacommons_services" {
   image                         = var.datacommons_services_config.image
   cpu                           = var.datacommons_services_config.cpu
   memory                        = var.datacommons_services_config.memory
+  cpu_idle                      = var.datacommons_services_config.cpu_idle
+  startup_cpu_boost             = var.datacommons_services_config.startup_cpu_boost
   min_instances                 = var.datacommons_services_config.min_instances
   max_instances                 = var.datacommons_services_config.max_instances
   make_public                   = var.datacommons_services_config.allow_unauthenticated_access

infra/dcp/modules/stack/variables.tf

@@ -45,6 +45,8 @@ variable "datacommons_services_config" {
     max_instances                   = number
     cpu                             = string
     memory                          = string
+    cpu_idle                        = optional(bool, false)
+    startup_cpu_boost               = optional(bool, true)
     google_analytics_tag            = string
     enable_mcp                      = bool
     search_scope                    = string
@@ -69,15 +71,17 @@ variable "auth_config" {
 
 variable "redis_config" {
   type = object({
-    enable                  = bool
-    instance_name           = string
-    memory_size_gb          = number
-    tier                    = string
-    location_id             = string
-    alternative_location_id = string
-    replica_count           = number
-    vpc_network_name        = string
-    vpc_connector_cidr      = string
+    enable                      = bool
+    instance_name               = string
+    memory_size_gb              = number
+    tier                        = string
+    location_id                 = string
+    alternative_location_id     = string
+    replica_count               = number
+    vpc_network_name            = string
+    vpc_connector_cidr          = string
+    vpc_connector_min_instances = number
+    vpc_connector_max_instances = number
   })
 }
 
@@ -106,6 +110,8 @@ variable "ingestion_config" {
     # Workflow & Helper Service
     workflow_lock_acquisition_timeout = number
     helper_service_image              = optional(string)
+    helper_service_cpu_idle           = optional(bool)
+    helper_service_startup_cpu_boost  = optional(bool)
 
     # Dataflow network configuration
     # Use WORKER_IP_PRIVATE when a compute.vmExternalIpAccess org policy

infra/dcp/variables.tf

@@ -151,6 +151,18 @@ variable "redis_vpc_connector_cidr" {
   default     = "10.13.0.0/28"
 }
 
+variable "redis_vpc_connector_min_instances" {
+  description = "Minimum number of VPC Access Connector instances. Lower values reduce baseline cost."
+  type        = number
+  default     = 2
+}
+
+variable "redis_vpc_connector_max_instances" {
+  description = "Maximum number of VPC Access Connector instances."
+  type        = number
+  default     = 10
+}
+
 # =============================================================================
 # Spanner Module
 # =============================================================================
@@ -279,6 +291,18 @@ variable "datacommons_services_memory" {
   default     = "16G"
 }
 
+variable "datacommons_services_cpu_idle" {
+  description = "When true, CPU is only allocated during request processing (cheaper for low-traffic services). When false, CPU is always allocated (better performance, avoids cold starts within running instances)."
+  type        = bool
+  default     = false
+}
+
+variable "datacommons_services_startup_cpu_boost" {
+  description = "Temporarily boost CPU allocation during container startup to reduce cold start latency."
+  type        = bool
+  default     = true
+}
+
 variable "datacommons_services_allow_unauthenticated_access" {
   description = "Allow unauthenticated access to the public-facing services of the Data Commons Platform"
   type        = bool
@@ -428,6 +452,18 @@ variable "ingestion_helper_service_image" {
   default     = null
 }
 
+variable "ingestion_helper_service_cpu_idle" {
+  description = "When true, CPU is only allocated during request processing (cheaper for low-traffic services). When false, CPU is always allocated."
+  type        = bool
+  default     = null
+}
+
+variable "ingestion_helper_service_startup_cpu_boost" {
+  description = "Temporarily boost CPU allocation during container startup to reduce cold start latency."
+  type        = bool
+  default     = null
+}
+
 # =============================================================================
 # Ingestion - Dataflow Network Configuration
 # =============================================================================
@@ -500,3 +536,10 @@ check "ingestion_dataflow_workers_limits" {
   }
 }
 
+check "redis_vpc_connector_instances_limits" {
+  assert {
+    condition     = var.redis_vpc_connector_max_instances >= var.redis_vpc_connector_min_instances
+    error_message = "The redis_vpc_connector_max_instances must be greater than or equal to redis_vpc_connector_min_instances."
+  }
+}
+