Backend – Integration and Deployment

1. Overview

This document outlines the integration and deployment procedures for the Kanto backend APIs and video processing servers across development and production environments. It defines how services are packaged, deployed, and configured in AWS and Google Cloud, including CI/CD practices, environment management, and rollback strategies.

2. Architecture Summary

Kanto's deployment involves a single core backend API, payment APIs, and video processing servers, deployed across two environments with different cloud providers:

Development: Hosted on Google Cloud Run (Docker-based) and Google Cloud Compute Engine (video processing).
Production: Hosted on AWS Elastic Beanstalk (deployed via Bitbucket Pipelines) and Google Cloud Compute Engine (video processing).

3. Environments

Environment	Service Type	Provider	Domain	Deployment Method	Notes
`dev`	API General	Google Cloud	api-sandbox.kanto.co	Manual script	Docker via Cloud Run
`dev`	Payments API	Google Cloud	payment-dev.kanto.co	Manual script	Docker via Cloud Run
`dev`	Video Processor	Google Cloud	api-videos-dev.kanto.co	Manual script	Python app using FFmpeg on GCE (e2)
`prod`	API General	AWS	api3.kanto.co	Bitbucket Pipeline	ZIP package on Beanstalk
`prod`	Payments API	Google Cloud	api-payment-prod.kanto.co	Manual script	Docker via Cloud Run
`prod`	Video Processor	Google Cloud	api-videos-prod.kanto.co	Manual script	Python app using FFmpeg on GCE (e2)

4. Deployment Process

4.1 Development (Google Cloud Run)

The development environment uses Docker to containerize the Django API.
Deployment flow:
1. Run the script ./deploy.sh

4.2 Production (AWS Elastic Beanstalk)

Production deployment is done via a Python ZIP package.
Manual deployment process:
1. Package the Django project into a ZIP with required dependencies.
2. Upload and deploy to Beanstalk via eb deploy.
3. Environment variables are managed in the Beanstalk configuration.

4.3 Payment APIs (Google Cloud Run)

Both development and production payment APIs are deployed using Docker on Google Cloud Run.