AnsweredAssumed Answered

Workflow Problem Detection

Question asked by William May on Apr 17, 2019
Latest reply on Apr 17, 2020 by Mostafa Helmy

Is anyone familiar with the Workflow Problem Detection feature? I've pasted the help info below, but I'm not clear on what to consider when picking the maximum workflow transition time. Should this account for the longest delay nodes and max amount of time I would expect something to verify, or is it only looking at nodes that are "trying" to transition, but haven't for a certain period of time? 


Setting Up a Workflow Problem Detection Schedule

Setting Up a Workflow Problem Detection Schedule

You can schedule a process the system initiates to detect whether a workflow process carried out by a workflow node has not completed within a particular amount of time. This helps you pinpoint workflow problems that may arise, especially in the case where you are creating and testing custom workflows before you deploy them. What constitutes an excessive amount of time a workflow is “stalled” is arbitrary. You must decide that based on your experience managing and implementing workflows.

For example, you may have determined from past experience that a workflow transition would typically complete within 10 minutes. Moreover, you may also conclude that if it does not there might be a problem that requires further investigation. In this scenario, you would therefore specify a maximum time limit of 10 minutes and schedule the detection process to run on an on-demand (run the process now) or a periodic basis. If the process detects that a workflow process has exceeded the 10-minute threshold, the system indicates that a workflow job is in an error state.

A workflow job is in an error state if an activity node process in the workflow has exceeded the time threshold. For example, a workflow job under the Jobs tab would have a workflow state value of “error” if any of it’s activity nodes were detected as exceeding the time threshold.

Although the error state indication is registered for a threshold violation detection, it is important to note that this does not indicate that a workflow has irrevocably malfunctioned. It may successfully complete without any intervention from you or it may successfully resume after you have “poked” the workflow. See How to Respond to a Stalled Workflow Error for more information on “poking” a stalled workflow.