Deploy Flask App on Kubernetes - 2

Yogendra Tamang — Sat, 08 Aug 2020 20:48:42 +0000

This is second part of this article, where we did following:

Created Simple Flask App
Built Docker Image of the App
Pushed the image to DockerHub.

In this post, I will show you how you can deploy your app in kubernetes cluster. If you do not have access to multinode Cluster, you can install single node cluster, like Minikube or Docker Desktop
For my ubuntu 18.04 machine, I followed this tutorial

Lets first understand few resources in Kubernetes. For simplicity, I am making simple Book-Shop Scenario here.

Imagine you work in a book shop. Customers come and ask for the book. Its your job to fetch the book requested. Lets say a customer comes and asks for Harry Porter book, you go to the fantasy novel section, find there are 3 copies, take one of them and sell it to the customer.
Now you have 2 copies remaining. However, your shop has a rule, every book must have three copies. That means, you need bring another copy from the stock and put that along those two other books.

1. Pod

In above scenario, each book is a Pod. Like the book "Harry Porter" has its own contents, pods also run their own application. Pods are instances which run actual application.

2. Deployment

In above scenario, it was a rule that there must be three copies of every book. This rule controlled desired state of number of books in the shop. Like this, we have deployment in Kubernetes which controls how many pods should be running.

3. Service

There were 3 copies of "Harry Porter". You could have taken any one of them (It does not matter since they are same copies). Simply put, "Harry Porter" means any of those three books. Services are similar like this. It identifies group of pods as single unit.

All resources in Kubernetes are created using yaml configurations. We usually define pod inside deployment so need two yaml files. Lets create them.

1_deployment.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: simple-flask-app
spec:
  replicas: 1
  selector:
    matchLabels:
      app: simple-flask-app
  template:
    metadata:
      labels:
        app: simple-flask-app
    spec:
      containers:
        - name: simple-flask-app
          image: yogen48/simple-flask-app
          ports:
            - containerPort: 8083

2.2_service
Lets create a service.

apiVersion: v1
kind: Service
metadata: 
  name: simple-flask-app
spec:
  selector:
    app: simple-flask-app
  type: ClusterIP
  ports:
    - name: http
      port: 9999
      targetPort: 8083

Now as per these yaml files, lets create resources in kubernetes

kubectl apply -f 1_deployment.yaml
kubectl apply -f 2_service.yaml

Now, lets check if the there are pods, deployments and services.

kubectl get all

Now, lets check if we got our app running or not by port forwarding the service.

kubectl port-forard svc/simple-flask-app 8888:9999

Now lets curl

curl localhost:8888

Deploy Flask App on Kubernetes - 1

Yogendra Tamang — Mon, 03 Aug 2020 09:12:49 +0000

Introduction

This is part one of two part series:
Here, I will discuss about creating simple flask app, containerizing it and pushing it to image repository DockerHub.

Prerequisites

You have docker installed
Familiarity with Docker commands and Dockerfile. You can go through this blog

Create Flask APP

Lets create requirements.txt files for our app.

Flask==1.1.2
gunicorn==20.0.4

Flask comes with one development server, but we will use production-ready server, like gunicorn.

Create src folder and app.py file inside that. Our App will have only two endpoints:

/ -- returns welcome message.
/names -- lists countries I visited :P

from flask import Flask, jsonify
app = Flask(__name__)
from flask import Flask, jsonify
app = Flask(__name__)

country_list = [
    'Nepal',
    'India',
    'Germany',
    'Netherlands'
]

@app.route('/')
def index():
    response_dict = {
        "message": "Welcome to FlaskApp"
    }
    return jsonify(response_dict)

@app.route('/names')
def countries():
    country_dict = {
        'countries': country_list,
        'total': len(country_list)
    }
    return jsonify(country_dict)


if __name__ == '__main__':
    app.run(debug=True, host='0.0.0.0', port=8083)

Run your app in debug mode

python src/app.py

Run your app from gunicorn

gunicorn --chdir src app:app -w 2 -bind 0.0.0.0:8083

Lets create deploy.sh, we will execute this directly

#! /bin/sh
gunicorn --chdir src app:app -w 2 -b 0.0.0.0:8083

(Make it executable: chmod +x deploy.sh)

You can browse http://localhost:8083/

Building Docker Image

Lets write Dockerfile to create docker image of our app.
Here I am going with python-3.7-slim-buster.

This is one of light weight and recommended images to use.
Follow this if you want to learn more about other python images.

FROM python:3.7-slim-buster

COPY requirements.txt .

RUN pip install -r requirements.txt

COPY . /app

WORKDIR /app

ENTRYPOINT [ "./deploy.sh" ]

Now run following command to build image

docker build -t simple-flask-app .
This will take some time depending on your network speed. Once completed you will see success message like this

Run your docker image

docker run -p 8888:8083 simple-flask-app
If you browse http://localhost:8888/names, you will see following output.

Now Push your image to DockerHub using following commands

$ docker tag simple-flask-app <dokcer-hub-account>/simple-flask-app

$ docker push <dokcer-hub-account>/simple-flask-app

If you are not logged in, log in first.

$ docker login

DEV Community: Yogendra Tamang