GCP Setup

This guide walks you through preparing your GCP environment for EZ-CDC deployment.

Overview

EZ-CDC workers run in your GCP project and need:

1. A GCP Project with required APIs enabled
2. A VPC network with a subnetwork
3. Network access to your source and sink databases

GCP Project Setup

Enable Required APIs

EZ-CDC requires the following APIs enabled in your project:

gcloud services enable \
  compute.googleapis.com \
  logging.googleapis.com \
  monitoring.googleapis.com \
  --project=YOUR_PROJECT_ID

API	Purpose
Compute Engine	Worker instances (MIG, instance templates)
Cloud Logging	Worker and daemon log collection
Cloud Monitoring	Custom metrics and health checks

Verify APIs

gcloud services list --enabled --project=YOUR_PROJECT_ID \
  --filter="name:(compute.googleapis.com OR logging.googleapis.com OR monitoring.googleapis.com)"

VPC Configuration

Standard Mode

For standard connectivity, workers get ephemeral external IPs for internet access.

# VPC network
resource "google_compute_network" "main" {
  name                    = "my-vpc"
  auto_create_subnetworks = false
}

# Subnetwork in the target region
resource "google_compute_subnetwork" "workers" {
  name          = "ezcdc-workers"
  ip_cidr_range = "10.0.1.0/24"
  region        = "us-central1"
  network       = google_compute_network.main.id
}

Cloud NAT Mode

For Cloud NAT mode, the subnetwork does not need external IPs — Cloud NAT handles all outbound traffic. Cloud NAT resources are created automatically by EZ-CDC during deployment.

# Same VPC and subnetwork setup -- Cloud NAT is added automatically
resource "google_compute_subnetwork" "workers" {
  name          = "ezcdc-workers"
  ip_cidr_range = "10.0.1.0/24"
  region        = "us-central1"
  network       = google_compute_network.main.id
  # No external IPs needed -- Cloud NAT handles egress
}

Network Access to Databases

Workers must be able to reach your databases from the subnetwork. Ensure firewall rules allow outbound traffic to:

Database	Port	Protocol
PostgreSQL (source)	5432	TCP
StarRocks MySQL (sink)	9030	TCP
StarRocks HTTP (sink)	8040	TCP

If your databases are in the same VPC, ensure internal firewall rules allow traffic. If they are in a different VPC or external, ensure appropriate routing and firewall rules exist.

Verification

Check APIs

gcloud services list --enabled --project=YOUR_PROJECT_ID | grep -E "compute|logging|monitoring"

Check VPC and Subnetwork

# List VPC networks
gcloud compute networks list --project=YOUR_PROJECT_ID

# List subnetworks in region
gcloud compute networks subnets list \
  --network=YOUR_VPC_NAME \
  --regions=us-central1 \
  --project=YOUR_PROJECT_ID

Test Database Connectivity

From a test VM in the same subnetwork:

# PostgreSQL
nc -zv your-postgres-host 5432

# StarRocks
nc -zv your-starrocks-host 9030
nc -zv your-starrocks-host 8040

Next Steps

With GCP configured:

1. Service Account & IAM - Understand permissions
2. Create Deployment - Launch workers