GCP Infrastructure from Scratch

The Challenge

When I joined Teester (now part of the Skeepers group), a rapidly growing startup in the video advice sector, I faced a significant challenge. The existing infrastructure had been built by developers directly on the Google Cloud Console, creating several critical issues:

• Security vulnerabilities: Many services were publicly exposed (GKE, Cloud SQL)
• Network architecture: Everything deployed on the default network
• No Infrastructure as Code: All changes made manually through console
• Scalability concerns: Architecture couldn't support the startup's rapid growth
• No automation: Manual deployment processes prone to errors

The mission was clear: rebuild the entire infrastructure from scratch while maintaining business operations.

The Solution

Phase 1: Infrastructure Design & Planning

I started by designing a completely new, secure, and scalable architecture:

• Custom VPC network with proper subnet segmentation
• Private GKE clusters with authorized networks
• Private Cloud SQL instances accessible only from within the VPC
• Bastion host for secure access to private resources
• Load balancer configuration for high availability
• Redis caching layer for improved performance

Phase 2: Infrastructure as Code Implementation

I developed comprehensive Terraform modules for:

• Network Module - VPC, subnets, firewall rules for secure network segmentation
• GKE Module - Private cluster configuration, node pools, security settings
• Database Module - Cloud SQL instances, backup configuration, security policies
• Load Balancer Module - SSL certificates, routing rules, health checks
• Monitoring Module - Cloud Monitoring, alerting policies, dashboards
• IAM Module - Service accounts, roles, permissions management

Phase 3: CI/CD Pipeline Setup

Working closely with the lead developer, I implemented:

• Dual deployment strategy: Applications deployed to both old and new environments
• Database synchronization: Real-time sync between environments
• Automated testing: Comprehensive test suite for infrastructure changes
• Rollback procedures: Quick recovery mechanisms in case of issues

Phase 4: Migration Execution

The migration was executed during a planned maintenance window:

1. Final database synchronization
2. DNS switchover to new infrastructure
3. Traffic routing through new load balancers
4. Monitoring and validation of all services
5. Decommissioning of old infrastructure

Technologies Used

• Google Cloud Platform: Primary cloud provider
• Terraform: Infrastructure as Code
• Kubernetes (GKE): Container orchestration
• Docker: Containerization
• GitLab CI: Continuous Integration/Deployment
• Cloud SQL: Managed PostgreSQL database
• Redis: In-memory caching
• Cloud Load Balancer: Traffic distribution

Results & Impact

✅ Security Enhancement

• All services moved to private networks
• Proper IAM roles and permissions implemented
• VPC firewall rules for network segmentation
• Encrypted connections throughout the stack

✅ Zero-Downtime Migration

• Successfully migrated without service interruption
• No data loss during the transition
• Seamless user experience maintained

✅ Improved Scalability

• Auto-scaling GKE clusters
• Database performance optimization
• Load balancer configuration for high availability
• Redis caching reducing database load

✅ Operational Excellence

• Infrastructure as Code for all resources
• Automated deployment pipelines

Client Testimonial

"Edwin worked with us for over a year on a complete overhaul of our cloud architecture and infrastructure modernization. Edwin is versatile: he's comfortable with the cloud, with Kubernetes and with pipeline management. Edwin brings valuable suggestions and helps us make the best decisions for our application deployment."

— Maxime Baconnais, Lead Tech at Teester

Project Duration: 6 months
Team Size: Solo DevOps Engineer + Lead Developer collaboration
Status: Successfully completed, infrastructure still running in production