code2exe / udacity-docker

Project IV for Udacity

Udacity Cloud DevOps Engineer Nanodegree, Project IV: Operationalize a Machine Learning Microservice API

This project tests the ability to apply the skills acquired in the nanodegree course to operationalize a Machine Learning Microservice API.

A model that has been pre-trained to predict housing prices in Boston according to several factors is given and the task is to operationalize/deploy it using containerization technologies like Docker and Kubernetes.

Steps:

• Setup the environment: (I am on Amazon Linux with python 3 installed)
  1. git clone https://github.com/code2exe/udacity-docker.git
  2. cd udacity-docker
  3. python3 -m venv ~/.udocker
  4. source ~/.udocker/bin/activate
• Running app.py: (Ensure that Docker and Kubectl with Minikube are properly installed)
  • Run with Docker: ./run_docker.sh
  • Run with Kubernetes:
    • minikube start
    • .\run_kubernetes.sh

Contents of the Repository:

• .circleci - CircleCI intergration for testing
• model_data- Contains pre-trained model data
• output_txt_files- Contains log output
• Dockerfile - Docker configuration
• Makefile- Makefile for local testing
• app.py - Our Flask app
• make_prediction.sh- Contains the payload to generate our prediction
• run_docker.sh - A script to automate our docker deployment
• run_kubernetes- A script to automate our Kubernetes deployment
• upload_docker- A script to build and push our Docker container to Dockerhub