Include "execution_time" or "execution_date" in context

[cicdw]

For flow runs, both local and cloud runs should include "execution_time" (or something similarly named) into context.

[cicdw]

And when this change occurs, update the examples/airflow_tutorial_dag.py file accordingly

[jlowin]

Per discussion: scheduled_start_time would be great. This would be the time a Flow Run was scheduled to start, as opposed to start_time, the time it actually entered a running state. Interestingly, start_time will not be immediately available (because it is only set when the flow enters a running state!) but users could always query for it if they needed it.

[cicdw]

scheduled_start_time is much trickier than I thought; my first reaction was to pull it from the Scheduled state that starts a Flow run. This works fine and well for the duration of that first run. However, if for some reason the flow run needs a second run through, I don't know where to pull this information from (as context won't persist into the second run, and the flow won't be starting with a Scheduled state anymore).

Similarly for start_time: if we place it in context when the flow run enters a Running state, but then it is submitted again due to retries, etc. when will be place this into context? And will it have a new value this second time around?

cc: @jlowin

[cicdw]

Also, for the record: things can be put in context via environment variable; so cloud could put it there and ensure it's there for each subsequent run. Although this would make local testing somewhat annoying.

[jlowin]

@cicdw pull a field on the flow_run itself, not its state, called scheduled_start_time:

query { 
	flow_run_by_pk {
		scheduled_start_time
    }
}