Google Cloud

Google Cloud Infrastructure

Overview

Selecting Google Cloud Platform (GCP) early on set the foundation for scalable, cost-effective deployment. As part of the AI-augmented design process, I analyzed cloud providers using ChatGPT, prioritizing: - Single-provider ecosystem** to minimize cross-platform complexity. - Minimal cost overhead** while maintaining flexibility. - Ease of configuration** to support DevOps best practices. Over time, Docker, GitHub, Kubernetes, and Terraform were introduced to enhance deployment, automation, and system management. These are covered separately based on when they were introduced.

Key Goals:

Status: In Progress

Complexity: Medium

Components

GCP & CLI

Deployed key cloud services, including MySQL, BigQuery, Redis, Cloud Run, and Artifact Registry.

SOARL Summary

    Situation:

    • Needed to select and deploy core tech stack components.

    • Chose MySQL for the database, BigQuery for billing-based reporting, and Redis for performance optimization.

    • Cloud Run** hosts Django, FAISS workers, and the public website, while Artifact Registry manages Docker images.

    Obstacle:

    • {“Cloud deployment required learning a range of new tools”=>”SSL Certificates, Docker, Configuration Management, Authentication, Cost Monitoring.”}

    Action:

    • Focused on low-cost, scalable deployment strategies, optimizing resource usage to keep cloud costs under $22/month.

    Result:

    • Gained proficiency in Google Cloud CLI, orchestrating deployments efficiently.

    • Built an environment that scales on demand while remaining cost-effective.

    Learning:

    • Hands-on cloud deployment demystifies cloud computing—breaking down GCP components made it easier to approach **AWS and Azure.

Key Learnings

Demos

Final Thoughts

During the first couple of months about the only thing I used was MySQL and a bit of Bigquery. I had to do minimal configuration as most of the time was spent either in Google Sheets, MySQL or Python and Django. Only once I transitioned to working on the website did I start to have to spend more time thinking about my GCP stack and focus on the various pieces. Each component on the stack was selected based on how easily and cheaply it could be deployed within GCP

Tags

Cloud Deployment DevOps Infrastructure as Code

Back to Portfolio