Deployment Parameters tab shows "Unset" for falsy deployment parameter values

[zzstoatzz]

This discussion was created from a Slack thread conversation.

Original Thread: https://prefect-community.slack.com/archives/C079VLLH5D3/p1761832347164709

---

Summary
In Prefect Cloud with Prefect 3.4.25, the Deployment > Parameters tab displays "Unset" when a deployment parameter is explicitly set to a falsy value (e.g., False), even though the parameter is stored correctly and used for runs.

Environment

• Prefect Cloud
• Prefect 3.4.25
• Deployments served locally (no work pool) using Flow.to_deployment(..., parameters=...) then Flow.serve(deployment)

Reproduction
Minimal example:

from prefect import serve, flow

@flow
def foo(flag: bool = True):
    return flag

if __name__ == "__main__":
    deployment = foo.to_deployment(name="foo", parameters={"flag": False})
    serve(deployment, pause_on_shutdown=False)

Steps:

1. Serve the deployment to Prefect Cloud
2. Open the deployment in the Cloud UI, go to the Parameters tab → shows "Unset" for flag
3. Run prefect deployment inspect "my_flow/repro" → shows parameters={"flag": false}
4. Start a run from the UI → the run uses flag=False (correct)

Expected behavior

• The Parameters tab should display the actual stored deployment parameter value (False) when a falsy value is explicitly set.

Actual behavior

• The Parameters tab renders "Unset" for falsy values, despite the parameter being present in the deployment's parameters dict and being applied at runtime.

Additional observations

• Parameter is top-level and boolean. Likely affects other falsy values (0, "", [], {}). Not tested with nested parameters.
• Behavior is purely visual; backend and runs behave correctly.

Workarounds

• Trust prefect deployment inspect and the pre-filled values in the Run drawer
• Temporarily set a truthy value to confirm visual rendering in the UI

Request

• Confirm and triage as a UI bug; update the Parameters tab rendering to correctly display falsy values that are explicitly set on the deployment.

Thanks!

---

This discussion was automatically created by the Marvin bot to preserve valuable community insights.

[m-lal]

Confirming that I saw that in my practical business case and also in the minimal example.