Project Overview

Udacity Cloud DevOps Engineer Nanodegree Project 5: Operationalize a Machine Learning Microservice API.

The sample app: A pre-trained, sklearn model that predicts housing prices in Boston according to several features: e.g. average rooms in a home, data about highway access, teacher-to-pupil ratios, etc. The project tests the ability to operationalize a Python flask app (app.py) that outputs predictions (inference) about housing prices through API calls.

Read more about the data, initially taken from Kaggle, on the data source site. This project could be extended to any pre-trained machine learning model, such as those for image recognition and data labeling.

Project Task Summary

1. Run project locally via python, and run code linters to check syntax:

• Install Python3.7
• Setup python venv

  python3 -m venv ~/.devops
  source ~/.devops/bin/activate

• Run make install
• Run make lint

2. Run project on a local Docker container:

• Run ./run_docker.sh
• Run ./make_prediction.sh
• Expected Console Output:

  Port: 8000
  {
    "prediction": [
    20.35373177134412
    ]
  }

3. Run project on a local kubernetes cluster (minikube):

• Download and install Minikube
• Run minikube
• Run ./run_kubernetes.sh

You may have to run the script again in order to wait for the pod to finish the creation cycle.

• Run ./make_prediction.sh
• Expected console output:

Forwarding from 127.0.0.1:8000 -> 80
Forwarding from [::1]:8000 -> 80
Handling connection for 8000