09/19/2022, 8:06 PM
Using kedro in databricks workflow job
I have a databricks workflow job using kedro import statement, that is showing as success even though the code has failed. I have a reproducible simplified example here. Seems like the failure error code is not being transferred from kedro to the databricks job.
# This import statement causes the workflow job to succeed, even though failures are reported. If I comment this line, errors are reported and job fails as well, which is the expected output.
from kedro.framework import session

except Exception as e:
I have attached the screenshot of how using the import statement changes the way databricks displays the error. By using Kedro import, it seems the error is confined within Kedro and not really sent to the calling databricks notebook? Hope this gives us some clue to what is going on.
I guess this is fixed in the latest release - - Rich traceback handling is disabled on Databricks so that exceptions now halt execution as expected.