AWS-Pulumi-MLflow-Remote-Tracking-Server

This is a README.md file for the the AWS Pulumi Stack for the creating MLflow Remote Tracking Server.

This code repository has all the code needed to create the AWS Pulumi Stack for the MLflow Remote Tracking Server and examples of how to use it.

Blogpost: https://mlops.community/mlflow-on-aws-with-pulumi-a-step-by-step-guide/

Prerequisites

• Install Pulumi
• Install Python
• Install AWS CLI
• Configure AWS CLI with your credentials (see here)

Cloud infrastructure

Deploy the cloud infrastructure

• Clone this repository
• Open a terminal and navigate to the folder src/cloud_infrastructure/
• Run pulumi up and follow the instructions
• Once the stack is deployed, you can check the resources created in the AWS Console

Update the cloud infrastructure

• Open a terminal and navigate to the folder src/cloud_infrastructure/
• Run pulumi up and follow the instructions

Destroy the cloud infrastructure

• Open a terminal and navigate to the folder src/cloud_infrastructure/
• Run pulumi destroy and follow the instructions

Automatic deployment of the cloud infrastructure

• The cloud infrastructure can be deployed automatically using Github Actions. To do so, you need to create a secret in the repository with the name PULUMI_ACCESS_TOKEN and the value the access token of your Pulumi account. Also you need to create a secret with the AWS credentials with the name AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY and AWS_REGION. Then, you can push the code to the repository and the cloud infrastructure will be deployed automatically for every push on the main branch.

Examples

To use the examples, you need to have the cloud infrastructure deployed and to have installed the requirements in the requirements.txt file, by running pip install -r requirements.txt.

Simple example

The file machine_learning_example.ipynb is an example of how to use the MLflow Remote Tracking Server for simple machine learning experiments. To run the example, you need to have the cloud infrastructure deployed, to have installed the requirements in the requirements.txt file and than you can run the notebook.

LLM example

The file llm_example.ipynb is an example of how to use the MLflow Remote Tracking Server for LLM evaluation experiments. To run the example, you need to have the cloud infrastructure deployed, to have installed the requirements in the requirements.txt file and than you can run the notebook.

Nginx configuration

The file nginx.conf is the Nginx configuration file for the MLflow Remote Tracking Server to use reverse proxy server and basic authentication. This file is used in the cloud infrastructure, to be precise in the EC2 instance.

Folder structure

.
├── .github                                         # Github actions
│   └── workflows                                   # Github actions workflows
│       └── deploy_cloud_infrastructure.yml         # Github actions workflow to deploy the cloud infrastructure
├── .vscode                                         # VSCode settings
│   └── settings.json                               # Settings for the formatting of the code
├── src                                             # Source code
│   ├── cloud_infrastructure                        # Pulumi stack for the cloud infrastructure
│   │   ├── services                                # Pulumi services
│   │   │   ├── ec2.py                              # Pulumi code for the EC2 service
│   │   │   ├── rds.py                              # Pulumi code for the RDS service
│   │   │   └── s3.py                               # Pulumi code for the S3 service
│   │   ├── __main__.py                             # Creation of the Pulumi stack
│   │   ├── Pulumi.dev.yaml                         # Pulumi configuration file for the development environment
│   │   ├── Pulumi.yaml                             # Pulumi configuration file
│   │   └── requirements.txt                        # Python Pulumi requirements
│   ├── examples                                    # Examples of how to use the MLflow Remote Tracking Server
│   │   ├── llm_example.py                          # Example of how to use the MLflow Remote Tracking Server for LLM evaluation experiments
│   │   ├── machine_learning_example.py             # Examples of how to use the MLflow Remote Tracking Server for simple machine learning experiments
│   │   └── requirements.txt                        # Python requirements for the examples
│   └── nginx_configuration                         # Nginx configuration files
│       └── nginx.conf                              # Nginx configuration file for the MLflow Remote Tracking Server to use reverse proxy server and basic authentication
├── .gitignore                                      # Git ignore file
└── README.md                                       # README file (this file)