Hello, Is there a way to make scheduled running workflow to be stopped if it's taking more than usual time? We have tons of workflows running as a daily schedule and sometimes one workflow gets stuck (don't know why) and it keeps trying to finish the process. Sometimes, it runs for over 12 hours overnight and all the workflows in queue after that gets stuck. It would be nice to have something that we can set to be "SKIP or STOP" after certain period of min/hour per workflow. Thank you.
Solved! Go to Solution.
Hi @mlgsleea,
You can configure the Server to stop long-running jobs in the Alteryx System Settings.
Navigate to the "Worker | General" page and enable the "Cancel jobs..." option as shown here, choosing a maximum runtime as appropriate for your workgroup:
If you have a particular job that is "hanging" the Server it's worth taking a closer look at the module itself and determining root-cause.
Best,
Steve A
Do note that this setting only applies to scheduled jobs.
Oh shoot. Is there a way to auto-kill manual jobs that run long? Those are the ones that can really cripple us if we aren't watching all the time.
Thanks.
Steve,
I saw a job that was manually executed on the Gallery that ran just over 4 hours even with our timeout setting at 7200 seconds.
Thoughts on how to troubleshoot this if it should work for scheduled and manually executed workflows?
Thanks!
Hi @npsmith -
That is indeed surprising behavior.
If you have access to the Server log files, perhaps you could dig a bit more deeply to see if we can get to root cause. If you don't have log files, I would suggest turning logging on as it's a good best-practice to facilitate diagnostic scenarios such as this.
If you could please search your log files for the moniker "AlteryxExecute_Run" and "AlteryxExecute_RunWizard", you should see one of these entries for each app that's been run on your system.
I tend to use Alteryx for such operations (see e.g., https://community.alteryx.com/t5/Engine-Works-Blog/A-deeper-dive-into-Server-logging/ba-p/22389). Conceptually you'll load your log files using a wildcard (*.log), blend and then use a filter tool to filter for the monikers plus the name of the app in question.
On the log line for executing the app, you should see a tag called "<Timeout>". If that tag is present, what is its value? And if it's not present then perhaps we just discovered root cause.
Thanks kindly for your patience.
Best,
Steve A
Fantastic setting, i use this quite often on server