Using "Run" in Server UI in conjunction with a scheduled flow

[ReubenTheDS]

Hi Team, I would like your inputs on using Runs triggered in Prefect Server UI when the Flow is a scheduled one.

Background:

1. I have a Prefect Server running on localhost:8080, along with a local agent & a project "ProjX" setup using the appropriate commands
2. The flow in my python script is scheduled to run everyday & "connects" to "ProjX" :

def get_todays_date():
    return f'{pendulum.now().to_date_string()}'

@task(log_stdout=True)
def task1(date):
    print(f'{date}')

@task(log_stdout=True)
def task2(res1):
    # ... other cool stuff with res1

schedule = IntervalSchedule( start_date=datetime.utcnow(), interval=timedelta(days=1))

with Flow("My_flow", schedule=schedule) as flow:
    date_to_print = Parameter("date_to_print", default=[get_todays_date()])
    res1 = task1(date_to_print)
    res2 = task2(res1)
    # more tasks here
    # and here

flow.register(project_name="ProjX")

I confirm I can see the scheduled flow-runs in the UI

3. In the UI, I can trigger a flow-run to print "2020-12-25" using:

{'date_to_print':'2020-12-25'}

The expectation:
Every day the scheduled flow runs, it should print the date for that day

The problem:
The day printed by scheduled flow-runs is the first day that I set up the Prefect Server & connected to it from the script.

What am I missing, or, how could I do this better?

[cicdw]

Hi @ReubenTheDS - this is a common confusion for users new to deferred computation.

On this line:

    date_to_print = Parameter("date_to_print", default=[get_todays_date()])

your get_todays_date function runs, and its return value is hardcoded as the Parameter's default. In your case, it might be easier to rely on datetime values stored in Prefect context than writing up custom logic.

[ReubenTheDS]

Thank you for the clarification . prefect.context helps