brontomerus / ml-workflows Goto Github PK

7.0 7.0 1.0 363 KB

A project to provide an easy way to get started building distributed environments for machine learning workflows in AWS using Prefect and Dask.

License: BSD 3-Clause "New" or "Revised" License

Python 76.41% Jupyter Notebook 23.29% Shell 0.30%

ml-workflows's People

Contributors

Stargazers

Watchers

ml-workflows's Issues

Agent Failure

The agents fail to build in the infrastructure pulumi code.

The current build fails pulling the container, indicating that it is an issue with the permissions or VPC networking

MFA

Add MFA onto the vpn-layer infrastructure code. This might require setting up an active directory, but will provide much more sincere security.

Load balancer typos

ml-workflows/infrastructure/loadbalancing-layer/__main__.py

Line 41 in 1bd3149

subnet_id=private_subnet_ids[0],

This variables private_subnet_ids does not exist.

It is very useful to SSH into the private subnet in order to discover what might be causing issues during development. By building this into the infrastructure code, it will make it exceedingly more easy to manipulate the internal ECS Clusters and Containers, as well as troubleshooting other issues.

Recommend Projects