helm-repo

Simple Chat App

Requirements

• Write a Chat service as Python.
• A simple html/js client that uses the service is provided (index.html)

1. What is a Chat service?
   A chat service accepts three APIs

2. Methods:

   a. GET /<room> - returns the static html provided to you, regardless of the room provided.
   b. GET /chat/<room> - returns the full chat in a room, data is load from file /data/<room>.txt and show the page
   c. POST /chat/<room> - accepts a chat line from a user

   • save to file /data/<room>.txt
   • accept 2 form fields - username & message
   • chat formatted as a list of “\n” delimited lines (1 per message)
   • each line formatted as “[2018-02-25 14:00:51] long.vu: hi everybody!”
   • save date, time, username & message per room

   <room> is the number such as 123 or 1234, etc.

3. Following is a recommendation on implementation order:
   a. Implement Flask app that returns the static html on GET /.
   b. Now add new page GET /<room> will show the same page above.
   c. Implement GET /chat/<room> will show the data loaded from file (create by manual for now).
   d. Implement POST /chat/<room> to save chat data. Check with another room and another browser.

4. DevOps Section
   a. Package application in docker-image
   b. Create docker-compose toplogy comprising of:

   • 3 chat servers
   • nginx container configured to load balance the 3 instances and expose them on port 80.
     NOTE that you do not need to expose ports for any of the 3 servers!
     You can read about nginx load balancing at
     https://docs.nginx.com/nginx/admin-guide/load-balancer/http-load-balancer/ and https://www.nginx.com/resources/wiki/start/topics/examples/loadbalanceexample/
     or just google it :-)

   c. Test and prove that:

   • You are accessing all 3 servers (The UI should show a friendly server name)
   • A chat message sent to one is visible on the others

   d. Make this work with arbitrary scaling, i.e. 1 nginx and several chat servers as defined by docker-compose “scale” parameter.
   Note that in the case of –scale, it is good enough if the UI just shows any indication that differs between servers. It does not have to be “friendly”

5. Bonus Section
   Replace the file based storage with a dockerized MySQL.

Implement as much as you can. Goodluck!!!

Nice to have all results

[Test with docker-container]

[Test with docker-compose]

[Test with docker-compose –scale]