Prices unknown for planning window.

[kuhlivisj]

When I try to trigger a schedule it always fails with the error "Prices unknown for planning window.".

I'm triggering the schedule with the API. I also tried the HEMS tutrial with the flexmeaures-client and there I get the same error.

The prices are available in my system:

I'm running the following version:

[Flix6x]

Please hit the replay button on your price sensor page to verify that the prices were known at 10 AM UTC. This is the starting time of the schedule, which corresponds to the default information cutoff used by the scheduler (encoded in the belief time). If the prices were recorded at a later time, the schedule won't take them into account, unless you specify a later belief time (the corresponding API field may be named "prior").

[kuhlivisj]

When I check the replay I see that the prices for the next day weren't available at 10 AM UTC

So am I sending the price data wrong to flexmeasures?

Everyday I send the prices for the next day to flexmeasures at the time I receive them.
For example today I used the following start and prior:

This was send to flexmeasures around 2026-01-22T13:30:00+0100

[Flix6x]

When I check the replay I see that the prices for the next day weren't available at 10 AM UTC

This shouldn't matter. Your original schedule started at 10 AM on the current day. At that time you already had prices for the rest of the current day, so the set of prices window the planning window should not be empty.

Everyday I send the prices for the next day to flexmeasures at the time I receive them.

This is good. Without a prior set they should be automatically registered as being recorded at the actual time FlexMeasures received them.

I'm trying to think of another reason, but if you could help debug that would be more fruitful. Would you be able to set breakpoints in your code and find out why the prices are not being loaded?

In flexmeasures/data/models/planning/storage.py this is the code block responsible for fetching the prices:

        # Check for known prices or price forecasts
        up_deviation_prices = get_continuous_series_sensor_or_quantity(
            variable_quantity=consumption_price,
            unit=self.flex_context["shared_currency_unit"] + "/MWh",
            query_window=(start, end),
            resolution=resolution,
            beliefs_before=belief_time,
            fill_sides=True,
        ).to_frame(name="event_value")

Under the hood it fetches the time series data in flexmeasures.data.models.planning.utils.get_series_from_quantity_or_sensor:

        bdf: tb.BeliefsDataFrame = TimedBelief.search(
            variable_quantity,
            event_starts_after=query_window[0],
            event_ends_before=query_window[1],
            resolution=resolution,
            # frequency=resolution,
            beliefs_before=beliefs_before,
            most_recent_beliefs_only=True,
            one_deterministic_belief_per_event=True,
        )

I suggest inspecting the arguments passed to TimedBelief.search when fetching beliefs from your consumption-price sensor.

[kuhlivisj]

Any tips on how I can debug?

I have downloaded the source and can run it in VSCode without the debugger, but if I try to start the debugger it doesn't run the code.

[Flix6x]

I'm not familiar with the VSCode debugger, but if you can run the app from source code and have a job ID of a failing job on the tasks page, you might try running flexmeasures jobs run-job --job <job ID> with a breakpoint() added to the code where you want to pause execution.