This overview discusses how to fix one workflow instance error at a time. If you find more than one error, follow all the steps for each instance. If one error requires a Relaunch, then fix all errors then Relaunch. See Cloudsnap help articles for details about how to complete specific tasks discussed here.
For a video about how to choose Resume or Relaunch in Cloudsnap, scroll to the bottom of the page.
Workflow errors are discovered two ways:
- Data check for accuracy and completeness at the destination.
- Notification from Cloudsnap that a workflow instance failed.
For details about failed workflow notifications and the types of restarts, see Restart Failed Workflow Instance.
Choosing how to restart a workflow depends on whether the data posted to the destination is correct. Whenever a workflow instance fails, fully investigate the error's cause before deciding.
Resuming restarts the process at the step before the error occurred, without doing anything to the records that were posted successfully. You cannot harm the results when you resume; however, the workflow will fail until the error(s) are fixed.
Relaunching reruns the instance from the trigger, which can create duplicates of records that processed correctly before the error occurred. Relaunch only as a last measure.
Mark as Success the failed workflow instance to remove it from the failed list in the Workflow Log before you relaunch the corrected dataset and/or workflow.
Decision Process for Resume or Relaunch
Scenario: Run a report from the destination database after the most recent instance of a workflow has run, but the data is incomplete or incorrect. This scenario always requires a Relaunch.
Search for the error(s) in the source data and the general workflow.
Repair all discovered errors whether they occurred in the source data or the workflow itself.
Mark as Success the failed workflow instance to remove it from the failed list.
Scenario: Automated notifications of failed workflow instances via email or Daily Digest. Find the failure in the Workflow Log.
Types of Errors
Before deciding how to fix the workflow, determine the type of error:
API / Function Error
Source Data incorrect or incomplete because did not meet API requirements.
Whenever an API Call is dropped due to an API disruption, check that the API has been restored, then RESUME.
The resolution for this type of error depends on where the workflow failed.
If the workflow failed at the step where the error occurred, fix the API or function in the workflow, then RESUME.
If the workflow failed at one or more steps after the actual error occurred, the input used in the step where the workflow failed was incorrect. That means that the step failed because the workflow attempted to process invalid data.
For example, if a step expects to process dates in one type of format but a previous step formatted dates in another, the current step fails because the dates are incorrect.
In this case, repair the API or function in the general workflow, then Mark as Success the failed workflow, and RELAUNCH.
Relaunching is necessary because the invalid data could be part of the input for other steps elsewhere in the workflow process.
Source Data Incorrect or Incomplete
When the source data did not meet the API requirements, most frequently, data is missing or incorrect for invoices, purchase orders, or vendor codes. Correct the data at the source then Mark as Success the failed workflow instance and RELAUNCH the workflow.