Because our System Flows can be running quiet sometime, sometime hours, and I don’t want my deployment killing a running system flow I have developed a way to check whether a system flow is still running.
What I do is:
- deactivated all running system flows
- check if there is a system flow running (is there an empty [end_date_time] in [process_flow_schedule_log])
- If so wait sometime and check again
- If not start deployment.
And of course I put on a max time the deployment will wait, if it takes more time, the deployment is aborted.
But now I’m seeing quite often that the [end_date_time] is not filled. It looks like that if a system flow encounters an error, it aborts and doesn’t write the [end_date_time] in it’s [process_flow_schedule_log].
Therefor my checks are failing and deployments are aborted much to often.
I would think that the [end_date_time] should always be filled.
Is there a solution for my problem or is this a flaw?
Best answer by Mark Jongeling
View original
