Has Anyone Experienced Random Workflow Failures with an OAuth Error?

Question

We recently updated to 2025.1, and we have a number of users that are now experiencing rolling workflow failures that they run from the Gallery. We're only seeing this error occur in Output Data tools that are writing to Snowflake, and in the case of workflows with multiple outputs to Snowflake, only one tool will fail while others write successfully. We've reviewed the configurations of the tools, ensured that credentials in DCM are tested and synced to the Gallery, and we're unable to find a resolution.

The error is always an OAuth Access Token Expired error. Sometimes after a failed scheduled run, a workflow will run successfully when manually run, other times it won't. Sometimes it will succeed for several scheduled runs in a row, error for a scheduled run, then go back to successfully executing on a schedule.

This was not an issue we experienced prior to the upgrade, I'm wondering if anyone might have insights into what we could investigate in our efforts to resolve this issue.

jdunleap · Answer

I'm going to let my updated workflows run for a few days to validate, but I used Control Containers instead of Block Until Done tools, and that did resolve a number of the broken workflows. Thank yo uso much!

jdunleap · Answer

That's a really good thought! I'm going to test that and I'll let you know what I find!

jacktorres4 · Answer

This sounds like a tricky race condition or a token lifecyle sync issue between Alteryx Server and DCM post-upgrade. Since workflows with multiple outputs sometimes only fail on one tool, it’s possible that concurrent tools are trying to refresh the same OAuth token simultaneously, causing one to invalidate the other. As a temporary workaround, have you tried adding a Block Until Done tool to sequence the Snowflake outputs so they don't hit the connection manager at the exact same millisecond? Defintiely worth opening a case with Alteryx Support as well