Markdown description can be hidden on pause_flow_run input box

[tunasplam]

Bug summary

There are some cases where the markdown description on the pause_flow_run input box can be hidden. It looks like it is an issue if there are too many inputs being requested.

Here is a simple example. Submit a flow with the first value as 0 and the pause_flow_run prompt will hide the error box. Reducing the number of input fields to 3 causes the input form to correctly render.

from typing import Self

from prefect import flow
from prefect.flow_runs import pause_flow_run
from prefect.input import RunInput
from prefect.states import Completed
from prefect.client.schemas.objects import State
from pydantic import Field, model_validator, ValidationError


class ToyExampleForm(RunInput):
    field_1: int = Field(description="Must not be 0", json_schema_extra=dict(position=1))
    field_2: int = Field(default=1, description="Field 2", json_schema_extra=dict(position=2))
    field_3: int = Field(default=1, description="Field 3", json_schema_extra=dict(position=3))
    field_4: int = Field(default=1, description="Field 4", json_schema_extra=dict(position=4))
    field_5: int = Field(default=1, description="Field 5", json_schema_extra=dict(position=5))
    field_6: int = Field(default=1, description="Field 6", json_schema_extra=dict(position=6))
    field_7: int = Field(default=1, description="Field 7", json_schema_extra=dict(position=7))
    field_8: int = Field(default=1, description="Field 8", json_schema_extra=dict(position=8))
    field_9: int = Field(default=1, description="Field 9", json_schema_extra=dict(position=9))
    field_10: int = Field(default=1, description="Field 10", json_schema_extra=dict(position=10))

    @model_validator(mode="after")
    def validate(self) -> Self:
        if self.field_1 == 0:
            raise ValueError("Field 1 cannot be 0")
        return self


@flow(validate_parameters=False)
async def deploy_toy_example(form: ToyExampleForm) -> State:
    validated_params = await validate_flow_inputs(form, ToyExampleForm)
    return Completed(message="👍")

DESCRIPTION_MD = """
    One or more of your inputs were invalid.
    Error Message:
    %s
"""

def _format_error(e: ValidationError) -> str:
    return "; ".join(err["msg"] for err in e.errors())

async def validate_flow_inputs(inputs: dict, form: type[RunInput]) -> RunInput:
    try:
        return form.model_validate(inputs)
    except ValidationError as e:
        msg = _format_error(e)

    while True:
        try:
            new_inputs = await pause_flow_run(
                timeout=3600,
                wait_for_input=form.with_initial_data(
                    description=DESCRIPTION_MD % msg,
                    **inputs
                )
            )
            return new_inputs
        except ValidationError as e:
            msg = _format_error(e)

Version info

Version:              3.6.24
API version:          0.8.4
Python version:       3.13.9
Git commit:           195d63c9
Built:                Fri, Mar 27, 2026 04:10 PM
OS/Arch:              linux/x86_64
Profile:              ephemeral
Server type:          cloud
Pydantic version:     2.12.5
Server:
  Database:           sqlite
  SQLite version:     3.50.4
Integrations:
  prefect-aws:        0.5.13
  prefect-dask:       0.3.6
  prefect-email:      0.4.2

Additional context

Here is a screenshot of the issue.

Here is a screenshot after I edited the markdown box to force it to have padding.