19913d85-f662-4518-a9a1-7adc1793dfd4.png){kind=small}
I have an asset-job that may very occasionally tak...
# ask-communityg
Giovanni Paolo
04/06/2023, 6:14 PMI have an asset-job that may very occasionally take longer to process than its scheduled frequency.
I know I can limit execution concurrency if using a QueuedRunCoordinator. However is it possible to build a schedule that simply skips if there's a current instance of the job being executed?
s
sandy
04/06/2023, 6:38 PMHi Giovanni - there's no built-in option for this, but you could implement this logic inside the decorated function for your schedule.
It might require a little code spelunking, but you could basically use
context.instance.get_run_recordsRunsFilterg
Giovanni Paolo
04/06/2023, 6:40 PMoh, thanks for the pointers! i'll do some digging
Giovanni Paolo
04/06/2023, 6:41 PMI was thinking maybe a combination of concurrency-limiting and should_run but I wasn't sure when should_run would be evaluated
Giovanni Paolo
04/06/2023, 6:41 PM(especifically to drop runs that had been queued for longer than some specified amount of time)
Giovanni Paolo
04/06/2023, 7:07 PMI think I managed to get it to work as expected. Thank you, Sandy!
here's a snippet for anyone stumbling into this in the future
Copy code
@schedule(job=my_job, cron_schedule="* * * * *")
def my_schedule(context: ScheduleEvaluationContext):
f = RunsFilter(job_name="my_job", statuses=[DagsterRunStatus.STARTED])
runs = context.instance.get_run_records(filters=f)
if len(runs):
return SkipReason(f"skipping because there are {len(runs)} started runs")
return RunRequest(run_key=None)s
sandy
04/06/2023, 8:47 PMawesome
2 Views