Great question. In Prefect 3.x you can set environment variables for your deployment via the work pool’s job variables. If your deployment will run on a Process work pool, add an

env

dict to

job_variables

like this:

flow.from_source(
    source=source,
    entrypoint="src/app/workflows/prefect_flows/mdm_mcp_ingestion_flow.py:mdm_mcp_ingestion_flow",
).to_deployment(
    name="mdm_mcp_ingestion_flow",
    description="Ingest MDM and MCP data into the database",
    tags=["etl", "mdm", "mcp", "ingestion"],
    version=version,
    work_pool_name=work_pool_name,
    job_variables={
        "env": {
            "PYTHONPATH": "src/app"
        },
        # optional: ensure the relative path resolves the way you expect
        # "working_dir": "/absolute/path/to/your/project"
    },
).apply()

Notes: - For the Process worker,

job_variables.env

maps to

ProcessVariables.env

, which is exactly for environment variables. - If you use a relative

PYTHONPATH

like

src/app

, it’s resolved relative to the process working directory. Either set an absolute path or also set

working_dir

in

job_variables

so it resolves correctly. - Alternative options: - Set env vars at the work pool level using a base job template so all deployments inherit them. - Export

PYTHONPATH

in the environment where you start the worker (e.g.,

export PYTHONPATH=src/app && prefect worker start ...

). Docs for more detail: - Customize job variables - Process worker variables Quick check: which work pool type are you using (Process, Docker, Kubernetes)? The exact

job_variables

schema differs per pool, but they all support an

env

section.