Start tracking the schedule that generated a flow run with the CreatedBy field

[abrookins]

We currently don't track which schedule created a flow run, but we'll need that information in the worker to determine if the worker should use a global concurrency limit.

We can use the CreatedBy column with a new value type instead of adding a new FK to flow_run. When the scheduler creates a scheduled flow run, it should capture the schedule ID like this:

created_by=CreatedBy(
  id=deployment_schedule.id,
  type="DEPLOYMENT_SCHEDULE",
  display_value="something something"
)