Cluster (Docker)

Cluster (Docker)

Reminder A Cluster is only an entity that can create and manage jobs.

It should has the following capabilities (To be continued):

1. Create a job from a Job Description
2. Get the job info (IP Address, status)
3. Get Logs
4. Kill

These actions should require some form of authentication, and the changes should be reflected in the database.

Docker

Our first cluster is Docker-based, meaning we will use the Docker CLI to spawn jobs locally. By doing so, we can create a solid base for future expansions.

To spawn jobs we will use python-on-whales. It looks simple enough for our usecase.