[GSK-680] Multi ML Worker support

[andreybavt]

🚀 Feature Request

Support multiple ML Worker connected at the same time

🔈 Motivation

To execute models trained using different python versions and different environments. To let different data scientists work with Giskard at the same time

🛰 Alternatives

Disconnecting an actively connected ML Worker and connecting a new one. This is a current solution but it's not great because it only allows a single active worker, so no way to run models trained in different environments at the same time.

📎 Additional context

In the settings admin should see all actively connected ML Workers
an active ML Worker should be assigned at a project level

❓Questions

Should workers have unique ids to detect if a worker that was selected as "active" for a given project isn't currently available?

_{From SyncLinear.com | GSK-680}