gcp-microservices-tf

Modular Terraform project for Google Cloud Platform (GCP) to host the Online Boutique app.

Architecture Implemented

• VPC: prod-vpc with routing_mode = "REGIONAL"
• Subnets:
  • gke-subnet (GKE nodes/pods/services)
  • serverless-subnet (Serverless VPC Access)
• Connectivity:
  • Private Services Access peering (for Cloud SQL + Memorystore)
  • Cloud Router + Cloud NAT for private GKE egress
  • Serverless VPC Access connector for Cloud Run
• Compute:
  • Regional GKE Standard cluster (private nodes, Workload Identity)
  • Optional Cloud Run frontend with all egress through VPC connector
• Data:
  • Cloud SQL PostgreSQL (private IP only)
  • Memorystore Redis Standard HA (private access)
• Ops:
  • Cloud Scheduler HTTP warm-up/health-check every 5 minutes
  • Persistent disk + Kubernetes StorageClass for stateful components

Project Structure

.
├── main.tf
├── outputs.tf
├── providers.tf
├── variables.tf
├── versions.tf
└── modules
		├── database
		│   ├── main.tf
		│   ├── outputs.tf
		│   └── variables.tf
		├── gke
		│   ├── main.tf
		│   ├── outputs.tf
		│   └── variables.tf
		├── network
		│   ├── main.tf
		│   ├── outputs.tf
		│   └── variables.tf
		└── serverless
				├── main.tf
				├── outputs.tf
				└── variables.tf

Required Inputs

Set these in terraform.tfvars (or via -var):

project_id   = "your-gcp-project-id"
region       = "us-central1"
cluster_name = "boutique-gke"

Optional:

disk_zone = "us-central1-b"
deploy_cloud_run_frontend = false
frontend_image = "gcr.io/google-samples/microservices-demo/frontend:v0.10.1"
allow_public_frontend = true
deploy_boutique_manifests = true
boutique_namespace = "boutique"
boutique_manifest_url = "https://raw.githubusercontent.com/GoogleCloudPlatform/microservices-demo/main/release/kubernetes-manifests.yaml"

Setup: Deploy Boutique Manifests via Terraform

You can deploy the Online Boutique Kubernetes manifests directly from Terraform.

Prerequisites on the machine running Terraform:

• gcloud
• kubectl
• gke-gcloud-auth-plugin

Why: Terraform uses a local-exec step to run kubectl apply after GKE is created.

Enable in terraform.tfvars:

deploy_cloud_run_frontend = false
deploy_boutique_manifests = true
boutique_namespace = "boutique"

Then run:

terraform init
terraform apply -var="project_id=your-gcp-project-id"

Usage

terraform init
terraform plan
terraform apply

Python Deploy Script (gcloud SDK)

You can deploy infra + app in one flow with:

python3 deploy.py \
	--project-id your-gcp-project-id \
	--region us-central1 \
	--cluster-name boutique-gke \
	--auto-approve

Useful flags:

• --skip-auth (if already authenticated)
• --skip-terraform (deploy/update app only)
• --skip-app (infra only)
• --tfvars-file path/to/terraform.tfvars

Outputs

• gke_cluster_endpoint
• cloud_run_url (null when deploy_cloud_run_frontend = false)
• boutique_frontend_external_ip
• boutique_frontend_url
• database_private_ip