ooeid / building-a-cicd-pipeline2

'Building a CI/CD pipeline' project for the 'DevOps Engineer for Microsoft Azure' nanodegree program from Udacity

By Rob Foster

Updated 08/11/2020

Introduction

This is my submission for the 'Building a CI/CD Pipeline' project as part of the 'DevOps Engineer for Microsoft Azure' nanodegree program from Udacity.

This project contains a python application that is designed to predict housing prices in Boston (I did not create the python app myself). This repo will enable you to:

• Deploy the app in Azure CloudShell
• Deploy the app as an Azure App Service

Any commits to the GitHub repo trigger automated code testing using GitHub Actions. A pipeline has been created in Azure DevOps, and the updated code is also automatically tested in Azure DevOps and deployed to the Azure App Service.

Here is an architectural diagram:

A Trello board has been created to keep track of tasks to be completed.

A spreadsheet has been created to manage the project schedule.

See here for a YouTube video demonstrating the project.

Instructions

Deploy the app in Azure Cloud Shell

In Azure Cloud Shell, clone the repo:

git clone [email protected]:bobfoster1299/building-a-cicd-pipeline2.git

Change into the new directory:

cd building-a-cicd-pipeline2

Create a virtual environment:

make setup

Activate the virtual environment:

source ~/.udacity-devops/bin/activate

Install dependencies in the virtual environment and run tests:

make all

Start the application in the local environment:

python app.py

Open a separate Cloud Shell and test that the app is working:

./make_prediction.sh

The output should match the below:

Deploy the app to an Azure App Service

Create an App Service in Azure. In this example the App Service is called rob-udacity-webapp and the resource group is called rob-udacity-project:

az webapp up -n rob-udacity-webapp -g rob-udacity-project

Next, create the pipeline in Azure DevOps. More information on this process can be found here. The basic steps to set up the pipeline are:

• Go to https://dev.azure.com and sign in.
• Create a new private project.
• Under Project Settings create a new service connection to Azure Resource Manager, scoped to your subscription and resource group.
• Create a new pipeline linked to your GitHub repo.

Screenshot of the App Service in Azure:

Screenshot of a successful run of the project in Azure Pipelines:

To test the app running in Azure App Service, edit line 28 of the make_predict_azure_app.sh script with the DNS name of your app. Then run the script:

./make_predict_azure_app.sh 

If it's working you should see the following output:

You can also visit the URL of the App Service via the browser and you should see the following page:

View the app logs:

az webapp log tail -g rob-udacity-project --name rob-udacity-webapp

Load testing

We can use locust to do a load test against our application. In this example we will do a load test against the app running locally rather than in Azure.

Install locust:

pip install locust

Ensure the app is running:

python app.py

Start locust:

locust

Open a browser and go to http://localhost:8089. Enter the total number of users to simulate, spawn rate, set the host to localhost:5000, and click Start Swarming:

You can then watch the load test:

Future improvements

Currently, there is only a single branch in GitHub. In the future it would be good to create multiple branches, so code can initially be tested and deployed in a staging environment. If it works correctly in the staging environment the changes could then be merged into the production branch and the code deployed into the production environment.