This is a simple web server which demos the use of Flask along with a load balancer. The environment is virtually created using Docker.

app runs at http://localhost/

mongo-express is available at http://localhost:8081

If you pass user_id as a query paramter in the GET request to /, the server responds with the associated user_key which it fetches from a mongodb database along with the port number of the service responding, otherwise it simply shows which Flask app is serving your request. The load is balanced within 3 flask servers using nginx.

Author: Shreyas Padhye

Machine should have Docker installed in it.

One-time setup:

1. Set mongo db credentials that you would want in two separate environment variables using following commands (where username and password are values that you want to set):
   a. export mongo_user=username
   b. export mongo_pass=password
   Note: sample commands specifically for mac & linux systems
2. Mount containers: Open root project directory
   run docker-compose up --build
3. SSH in mongo's docker container: Open a separate terminal in root project directory
   run docker exec -it mongo /bin/bash
4. Here, create database using dunp: In the docker container terminal from step 3,
   run mongorestore --db myusers dump/myusers -u username -p password --authenticationDatabase admin
5. Run following commands to create role and authentication for myusers database:

. mongo -u username -p password --authenticationDatabase admin
. use myusers  
. db.createUser({ user: "username", pwd: "password", roles: [{ role: "readWrite", db: "myusers" }] }) 

Note: use same username-password that you created in ENV variables

6. run exit to exit docker container shell
7. Proceed with testing (refer to Test section)

For every time you want to run / restart containers:

1. Open terminal in project's root directory and run following commands
2. docker-compose build
3. docker-compose up
   or alternatively: docker-compose up --build

To stop running services: Press Ctrl + Z and RUN docker-compose stop

On Browser:

• To test load balancer working, you can keep reloading the home page at http://localhost/ and you'll see the page being served from a different flask server everytime.
• To test user_key retrieval, pass in the user_id as a query parameter to '/' route eg http://localhost/?user_id=1 will return one@5cdb5896d7c8 as the key. The response will also contain the port of the flask server that is responding.
• If a user_id is passed which does not exist in the database, the server returns an error with 400 status code
• If the request does not contain the user_id param, it will return port
• user_id from 1 to 7 are available in database and more can be added through mongo-express

Additionally, you can use a REST API client like Postman to GET request http://localhost/ with the user_id query param

• docker-compose ps
• mongodump --db myusers -u username -p password --authenticationDatabase admin

1. Carefully went over the requirement and understood different aspects of the assignment.
2. Created virtual environment for flask app along with its folder structure and created requirements.txt file so that it can be served through uWSGI
3. Created wsgi config & basic flask routing at '/' and tested working through flask run
4. Dockerized flask and nginx
5. Added code to spin off 3 flask servers via docker containers
6. Added load balancer code to share load among the three flask servers
7. Added mongo integration and included dump restore functionality in Readme to ease deployment
8. Implemented query params handling to query mongodb and fetch appropraite key

 - Handle server side exceptions eg when Internal Server Error happens due to authentication failed
 - Change Dockerfile code to use ENV variable to set .ini file & implement through single docker file
 - Explore if it can be done through single app.ini file
 - Can add checks for DDoS and block IPs