The per-enrollment call to get progress data can fail. We need to handle it gracefully.
In the new 'workflow next' code, force failures on enrollment processing withing the course / enrollment loop and test for graceful handling
If an enrollment fails, check that the course is not bad / invalid
Track the failures. One option here is to add metadata to the figures.models.EnrollmentData record. This can help make it easier to query, inspect data, assess than to put into a special 'error record' like we've done with figures.models.PipelineError
We need to improve overall error tracking and provenance in the pipeline code anyway to help manage it at scale
The per-enrollment call to get progress data can fail. We need to handle it gracefully.
In the new 'workflow next' code, force failures on enrollment processing withing the course / enrollment loop and test for graceful handling
figures.models.EnrollmentData
record. This can help make it easier to query, inspect data, assess than to put into a special 'error record' like we've done withfigures.models.PipelineError