-
Notifications
You must be signed in to change notification settings - Fork 530
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add Mlflow 403 PL UserError #1623
Conversation
Co-authored-by: Daniel King <43149077+dakinggg@users.noreply.github.com>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Can we also add a try/catch to the read operation when we download data via uc.files.download(...)
? @mattyding
I thought motivation for the PFC->wrapping change is that we have no examples of customer runs with data read errors and can't validate the PFC check. So any try/catch wrapping we do for read would be a best guess. I say we proceed with just the 403 artifacts error for now because we have ample evidence for that. If another data read error comes up (still waiting on Rohit's logs), then can add another wrap |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
lgtm
Co-authored-by: Daniel King <43149077+dakinggg@users.noreply.github.com>
Co-authored-by: Daniel King <43149077+dakinggg@users.noreply.github.com>
This PR
Adds networking UserError to Foundry.
Abandoning PFC change in favor of this. This try-catch should sufficiently cover the 403 networking errors we are seeing. Additionally, change is safer as it avoids duplicating logic in composer.
Testing
Was gonna add a unit test that mocked a MLflow 500 exception but the test doesn't really fit anywhere in the testing suite and has strong assumptions, so not much benefit. Idk if others have thoughts
Rollout (including migrations)
DLE release