Deploy

Deployment Workflow Documentation

file:./deploy.png

Overview

This document outlines the deployment workflow of a server application involving a migration. The workflow consists of three main stages:

1. Local Development
2. Development Environment on GCP
3. Production Environment on GCP

Local Development

• Components: Docker container for PostgreSQL, Docker container for App Server
• Responsibilities: Active development, running migrations locally
• Debug Pointers: Inspect Docker logs, run manual tests on local machine

Steps

1. Active Development: Write and modify code.
2. Docker Compose: Use docker-compose up -d to bring up Postgres and App containers.
3. Local Migrations: Run Diesel migrations against the local Postgres container.
4. Testing: Execute local test suite.

Development Environment

• Infrastructure: Dedicated GCP VM
• Components: Docker Registry, PostgreSQL on VM, App Server on VM
• Responsibilities: Automated deployment and testing
• Debug Pointers: Check CI/CD pipeline logs, verify Docker registry uploads, examine VM logs

Steps

1. Git Workflow: Push changes to a feature branch.
2. CI/CD: Automatically build and upload Docker image to registry.
3. Dev Migrations: Run Diesel migrations on Dev PostgreSQL.
4. Dev Deployment: Deploy new image as a Docker container on the Dev VM.
5. Automated Tests: Run any tests and report back.

Production Environment

• Infrastructure: Same as Development, but separated for isolation.
• Components: PostgreSQL, App Server
• Responsibilities: Final deployment, migration, and validation
• Debug Pointers: Check logs in the production environment, especially during and after migrations.

Steps

1. App Shutdown: Stop the production App container but leave Postgres running.
2. Prod Migrations: Execute Diesel migrations on production database through a temporary container.
3. Prod Deployment: Bring up new App container from updated Docker image.
4. Validation: Run production-specific tests to ensure everything is operational.

Key Takeaways

• For local debugging, consult Docker logs and local test outputs.
• In Dev and Prod, debugging involves checking pipeline logs, container statuses, and database logs.
• Migrations are executed against respective databases at the Dev and Prod stages.
• Testing is woven into each stage to ensure functional integrity.