local mlops monitoring

Inspiration

I want to setup monitoring for any ml project easily so I made this project.

What it does

It deploy whole monitoring setup using grafana and postgres, we can setup workflow orchestration and perform metric monitoring using evidently. We can use this as a template to setup mlops monitoring flow for any project.

How we built it

• Run docker-compose-up in root directory
• Open new terminal, create a virtual environment and follow the steps I mentioned in the project readme to run this project. Project Readme

Challenges we ran into

You should have python 3.11 installed for the virtual environment setup.

Accomplishments that we're proud of

You can easily make ml monitoring project and setup and share it locally using ngrok or lifecycle docker extension.

What we learned

How to use workflow orchestration tool with grafana and evidently for mlops monitoring.

What's next for local mlops monitoring

I will add unit testing using pytest Add integration testing using localstack and docker-compose Add mlflow for experiment tracking and use minIO docker container for s3 bucket. Create a cli template of this project using cookiecutter.

Built With

• docker
• docker-compose
• evidently
• grafana
• postgresql
• prefect