Deploy Flask Hello World Application on EKS

Automate Build using Jenkins

tree .

Directory structure

.
├── Dockerfile          Docker file for the container
├── hello.py            Hello World python file
├── README.md           This README file you are reading right now
└── shell-commands.sh   A list of useful commands

Step 0: Start and Configure your environment

Create Amazon Linux ec2 instance
install docker
install git
install jenkins
start docker
give current user and jenkins to use docker without root privelages

sudo groupadd docker # create group
sudo usermod -aG docker ${USER} # add currrent user
sudo usermod -aG docker jenkins # add jenkins

logout and login from ec2 user or Better restart your instance to apply changes
Test your Flask Application on the host machine (optional)

git clone https://github.com/anandshivam44/flask-hello-world.git
cd flask-hello-world
sudo yum install python3-pip
pip3 install Flask
python3 hello.py

Open ec2 instance IP in a new tab with port 8081 If it does not open check all port and check all security groups `

Step 1: Create/Analyse Dockerfile

# pick ubuntu image from Docker Hub
FROM ubuntu:20.04

# update list of available package && install pip3 package manager, dev essentials
RUN apt-get update && apt-get install -y python3-pip python-dev

# set work directory inside Docker image
WORKDIR /app

# install Flask inside container
RUN pip3 install Flask

# copy project files to WORKDIR
COPY . /app

# App started on port 8081 so expose the same port
EXPOSE 8081

# entrypoint for docker run command
ENTRYPOINT [ "python3" ]
CMD [ "hello.py" ]

Step 2.x is optional and for testing only

Step 2.0: Build Docker Image from Dockerfile manually

docker build -t flask-tutorial:v1 .

or directly pull source from this repository and build an image

docker build github.com/anandshivam44/flask-hello-world

Step 2.1: Run Image and test it in your local system

Enter this command and log in to your hub.docker.com credentials

docker run -d -p 8081:8081 flask-hello-world

The container runs in https://localhost:8081

Step 2.2: Login to DockerHub

Enter this command and log in to your hub.docker.com credentials

docker login

Step 2.3: Commit your container and push to docker hub

Copy ID of your Docker Container in the clipboard from

docker ps

Replace 68339d202950 with your container ID and commit your changes

docker commit 68339d202950 anandshivam44/flask-hello-world:v1

Push to Docker Hub. Replace anandshivam44 with your id, flask-hello-world with your container name and v1 with your tag name. Replace similar reference in this article.

docker push anandshivam44/flask-hello-world:v1

Step 3: Jenkins: Build container on new commit from Github and push to Docker Hub

Open Jenkins on ec2 instance. Open public ip on port 8080
First install Docker Plugins Go to Manage Jenkins > Plugins > Available Search Docker and install Docker plugin by CloudBees (CloudBees is the maintainer of the Jenkins project) Install Docker Plugin
Go to HomePage
Create a new job
Create a new Freestyle Project. Click OK
In the configuration page add Github URL https://github.com/anandshivam44/flask-hello-world
In Source code management select git add repository git URL https://github.com/anandshivam44/flask-hello-world.git
Since this repo is public no need to give Credentials
In Build Triggers choose Poll SCM. Add cron-job task "* * * * *" without quotes. This will check the Github repo every minute for any change. If there is a change it will start a new build
In Build section > Add Build Step > Choose 'Docker Build and Publish'
Add Docker Hub repository. In my case anandshivam44/flask-hello-world. The format is username/image-name
Choose/Add docker hub credentials
Save change
Click Build now

Step 4: Create Cluster in Amazon ECS

Create a cluster in Amazon ECS

Goto AWS ECS > Cluster and click on Create Cluster
Select cluster template as Networking only. Click NEXT.
Type a cluster name. Do not create VPC for demo easy deployment. Leave all to default and click NEXT.

Step 5: Create and Define Task

Goto Task Definition. Choose AWS Fargate. Click Next.
Type a Task Definition name of your choice
Select Task role to None. We don't need a role for now
Set Task memory to 1 GB and CPY to 0.25 vCPU
Click Add container definition
Type a container name that points to docker hub anandshivam44/flask-hello-world
Set container image to https://hub.docker.com/repository/docker/anandshivam44/flask-hello-world It will download from Docker Hub

anandshivam44/flask-hello-world

Set Soft limit to 300 MB
Set Port Mapping to 8081
Leave everything to default and click Add
Click create to create Task Definition

Step 6: Create Load Balancer

In a new Tab GoTo ec2 > Load Balancer > Create Load Balancer
Select HTTP/HTTPS Application Load Balancer
Name you Load Balancer your choice
Scheme - internet-facing
IP address type ipv4
Listen HTTP on port 80
Select default vpc (Mumbai) and choose subnet 1a and 1b. Do not choose 1c. 1c does not support t2 micro free instance
Leave all defaults and click Next
In Security Settings leave everything as it is and click next
Create a new security group. Name it. Open TCP on port 80. Click Next. Protocol HTTP. Port 80.
Create a Target group. Target type IP. Leave all default. Click Next
Leave Register Target Type as empty
Review and create LB
Wait for ABS to create before you proceed. It will take some minutes.

Step 7: Create Service

Go back to ECS > Clusters > Choose your cluster > Services
Click create a cluster
Choose launch type AWS Fargate
Choose the task definition that we created
Give the service a name
Number of tasks as 2 (You can even choose 1 or n)
Leave all default and click Next
Choose the VPC as default as we selected before. Subnet Mumbai 1a and 1b
Choose a security group for Service
Auto-assign IP Enabled. Leave al default. Click Next
Choose Application Load Balancer and select the one we created
Click 'Add to load balancer' IMPORTANT
port 80 HTTP
Choose the Target Group we created
Leave all default. Click Next
Leave Auto Scaling to default. Click Next
Click Create service

Step 8: Verify Security Groups

Till now one Security Group (SG) was assigned to our ECS Cluster and one to our ECS ec2 instance. Goto SG and allow LB to talk to Cluster ec2. In inbound rules allow all TCP from LB Security Group to ECS ec2 so that health check succeeds.

Step 9: Hurray! Test you connection

Goto LB. Copy its DNS URL. and test your connection. You should see a hello world on your screen

Step 10: Cleaning up

Delete ec2 instance
Goto Dockerhub and delete your repository
Goto AWS ECS and Delete Cluster and Task
Goto AWS ALB and delete LB

anandshivam44 / flask-hello-world