Hello! I think it could be useful to add message outputs to the schedule

should_execute

function. For example, I check for various conditions such as if certain environment variables are set before launching a pipeline, and if a run was skipped it sometimes takes some time for me to find out exactly why. Adding a mechanism for outputting messages here would help debug why runs were skipped.

Hey Leo. How are you constructing your schedule (e.g. with a decorator, or by defining a

ScheduleDefinition

? The

ScheduleDefinition

class takes in an

execution_fn

argument that has the same signature as a sensor evaluation function. You can directly yield

RunRequest

and

SkipReason

objects from that function. https://docs.dagster.io/_apidocs/schedules-sensors#dagster.ScheduleDefinition

I am currently using the

schedule

decorator. Is there something similar I can use for that? If not, I think I should be able to use the

ScheduleDefinition

instead.

No, unfortunately not… I’ve been meaning to deprecate the run_config_fn based

@schedule

in favor of the execution_fn based decorator and keeping the decorator name to

schedule

but that’s a breaking change and hadn’t figured out a reasonable transition plan.

Sounds good! I will move away from the current

@schedule

in that case.

@Leo Kell Just wanted to follow up… we just released

0.13.10

which now supports using the

@schedule

decorator with the

RunRequest

based evaluation function. You can switch back to using that instead of manually constructing the

ScheduleDefinition

. Something like this:

@schedule(cron_schedule="* * * * *", job=my_job)
def my_schedule(context):
    if os.getenv("SHOULD_SKIP"):
        yield SkipReason("skipping")
        return 
    else:
        yield RunRequest(run_config=..., tags=...)
        return

That's great! Thank you for the update.