Hi all I have setup an airgapped prefect cluster and I'm now trying to work up a strategy for deploying flows. I am required to use work pools of type

process

and since the cluster cannot access the Internet, I'm also storing flows locally on each node. So I have setup a couple of nodes, each running

prefect worker

with a work pool of type

process

. I have successfully came up with an ansible playbook which deploys my sample flow to the cluster. However, I would like to be able to use an isolated environment for each flow - how can I achieve this when using the

process

work pool type? I have already created a venv for my flow and would now like to configure the prefect deployment with instructions to use it. This is my

prefect.yaml

file:

deployments:
  - name: "my_workflow"
    entrypoint: "main.py:my_workflow"
    work_pool:
      name: green
      work_queue_name: low
    job_variables:
      command: |
        echo "About to activate" && . /opt/prefect/flows/demo1/activate.sh && /opt/prefect/flows/demo1/env/bin/python -m prefect.engine
      working_dir: "/opt/prefect/flows/demo1"

unfortunately this does not seem to work - the flow is deployed successfully and I can trigger workflow runs just fine. However they do not use the custom environment, but rather run with the same environment being used by the prefect worker process. I was hoping to be able to override the python env with that

job_variables.command

property of the deployment, but it seems to be silently ignored. Is there some other way to do this, maybe using env variables? Thanks for helping out!