I just shipped my new backlog software — and this wasn’t just a weekend project:
- Over 195,000 words (194641 to be exact) of engineering, research, and problem-solving (including nearly 34,000 words (33882) with Gemini 2.5 Pro as my AI copilot) went into the full stack.
From Day One, it was Infrastructure as Code:
- 🌎 Terraform powered the entire Google Cloud foundation:
Projects, Cloud Run, SQL (Postgres), Artifact Registry, service accounts, networking, and permissions — all version-controlled, reproducible, and audit-proof.
What’s live and running in production?
- ☁️ Google Cloud Run: Fully serverless Python/Flask app, scalable and maintenance-free.
- 🗄️ Cloud SQL (Postgres): Secure, reliable database in the cloud — no manual setup.
- 🔒 Real user authentication: Credentials are hashed, managed, and validated in the database — no plaintext, no shortcuts.
- 🤖 Automated CI/CD with GitHub Actions:
- Every commit triggers a full build, container push to Artifact Registry, and automatic deploy to Cloud Run.
- Uses secure key-based authentication via GitHub Secrets and Google Cloud IAM Service Accounts (WIF/OIDC coming soon).
- 📦 Artifact Registry: Modern, managed container storage; no more legacy
gcr.io. - 🛠️ Zero manual steps: Every resource, permission, and deployment is code-driven.
- 🔑 Secret management: All sensitive info managed as GitHub Actions secrets — never in code or config files.
Takeaways:
- Terraform-first made it fast to scale, test, and iterate — and future-proofs the whole stack.
- With a real CI/CD pipeline, every deployment is safe, repeatable, and production-ready.
- AI support is powerful — but real-world debugging, persistence, and community wisdom made the difference.
Next steps:
Feature upgrades, UI/UX polish, and maybe open source.
Infrastructure is code, automation is king — and everything is built to grow.
#Terraform #IaC #GoogleCloud #CloudRun #CI_CD #DevOps #Python #Docker #ArtifactRegistry #Automation #Flask #MVP #AI #Productivity