animogm / building-ci-cd-pipelines

Overview

This project Setup is on a Continuous Integration/Continuous delivery pipeline for a Python-based machine learning application using the Flask web framework.

Project Plan

• Trello Board
• Spreadsheet

Instructions

• Architectural Diagram (Shows how key parts of the system work)

Launch Azure Cloud Shell

• login to azure portal at https://portal.azure.com/ and open the cloud shell

Generate SSH key and upload the public key to github

ssh-keygen -t rsa

NOTE: use cat ~/.ssh/id_rsa.pub if you have generated it before

Clone the repository using Azure Cloud Shell

git clone [email protected]:AnimogM/Building-CI-CD-Pipelines.git

Create a new python virtual environnement

cd Building-CI-CD-Pipelines

python3 -m venv venv

source venv/bin/activate

Install dependencies and build the app

make all

Setup Github Actions

Create Azure App Service

az webapp up --sku F1 -n maryam-ml-app

Setup Azure DevOps

• Go to https://dev.azure.com and sign in.
• Create a new private project.
• Under Project Settings create a new service connection and select Azure Resource Manager.
• Create a Python-specific pipeline to deploy to App Service, and linked it to your GitHub repo.

Get app logs

az webapp log tail

Test the app is up and running

• set the app name inside make_predict_azure_app.sh file

• Output of streamed log files from deployed application

Enhancements

It an be improvedon by having multiple environments (Dev, Staging and Production) and have a pipeline for each environment.

Demo

https://youtu.be/S4hdH8pHzoA