Cloud Architecture & Cost Optimization

Objective & Constraints

The platform was hemorrhaging $4,200/month on managed database services from a third-party provider while simultaneously suffering from environment drift. Development, staging, and production databases had diverged in schema versions, causing subtle bugs that only surfaced after deployment. The lack of infrastructure as code meant every environment change was a manual, error-prone operation.

Strategic Implementation

Led DevOps efforts on GCP, overseeing deployments and monitoring while maintaining end-to-end reliability of the infrastructure. I migrated the managed database to a self-hosted, highly available setup within GKE to eliminate external vendor costs and implemented GitOps for deterministic deployments.

Protocol Execution

• Orchestration: Architected and maintained a highly available Kubernetes cluster on GKE using Terraform for reproducible infrastructure provisioning.
• Infrastructure: Managed full infrastructure using GKE and Bitbucket for automated flows, enforcing infrastructure-as-code principles across all environments.
• Reliability: Implemented comprehensive monitoring and alerting using Prometheus and Grafana for platform stability, providing granular visibility into cluster health.
• Optimization: Reduced external DB costs by standardizing deployments and migrating databases into StatefulSets with persistent volumes, backed by automated snapshot policies.

Professional Reflection

"Migrating from a fully managed database to a self-hosted stateful workload on Kubernetes is often seen as taboo, but with careful planning, robust storage classes, and automated backups, the cost savings justified the increased operational responsibility."