Akshaya DevOps Portal¶
Operational knowledge base for AkshayaBazaar: a React 18 + Vite frontend, a .NET 8 Clean Architecture backend, MySQL 8.2, Redis 7.2, deployed via Docker Compose (local), k3d/K3s + Helm (local staging and the real staging/production clusters), with infrastructure provisioned by Terraform (Hetzner Cloud + Cloudflare).
This is a step-by-step operational portal, not a code-reference dump - every page in Docker, Kubernetes & Helm, Terraform, CI/CD, Deployment, and Operations follows the same shape: Purpose → Prerequisites → Commands → Validation → Troubleshooting → Rollback, so you can go from "I need to do X" to a working result without guessing.
Current validated status¶
| Layer | Status |
|---|---|
master branch |
healthy - PR #7 (DevOps automation) and PR #4 (multi-seller order split) both merged |
| Local Kubernetes staging | .\scripts\verify-staging.ps1 → 14 PASS, 0 WARNING, 0 FAIL |
| Backend CI | Restore → Build → Test (xUnit) |
| Frontend CI | npm ci → npm run lint → npm run build |
Where to start¶
- New to the project, setting up your own machine? → Local Development
- Need to build/run a container? → Docker
- Deploying to the local k3d cluster or a real cluster? → Kubernetes & Helm
- Provisioning/changing cloud infrastructure? → Terraform & Infrastructure
- Understanding what CI/CD actually does? → CI/CD
- Running a deployment end to end? → Deployment
- Something's broken and you need to check/fix it now? → Operations
- Want the big picture first? → Overview & Architecture
Conventions used throughout this portal¶
- PowerShell is the default for anything run from a Windows developer workstation. Commands
that must run in Linux/Bash (Jenkins agents,
terraform/helminside a Linux container, CI) are explicitly labeled Linux/Bash. - Every command is copy/paste-ready. Placeholders are angle-bracketed and named for what they
are, e.g.
<MYSQL_PASSWORD>,<CLOUDFLARE_API_TOKEN>,<RAZORPAY_KEY_SECRET>- never a real value. - Read-only vs. mutating commands are always distinguished. A page that includes both labels each clearly - see Operations → Troubleshooting in particular.
- Production is never one copy/paste away from staging. Every deployment runbook keeps Staging and Production as separate pages with separate preconditions, not toggles on the same page.
Earlier working notes¶
This portal supersedes but does not delete the original DevOps working notes written during the infrastructure build-out - kept in full under Legacy Reference for the detailed history (exact audit findings, live-cluster debugging transcripts) behind decisions summarized throughout the pages above.