-
Notifications
You must be signed in to change notification settings - Fork 1.6k
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
[Front end] [FR] Better error message for custom metric viz. #3499
Comments
/assign @Bobgy This is not urgent but assign to you to triage :) |
@numerology Thanks for report! Can you point me to where the regex is? |
https://www.kubeflow.org/docs/pipelines/sdk/pipelines-metrics/#export-the-metrics-file name must follow the pattern ^a-z?$. |
@thebinaryone1 Thanks! |
bumped priority |
See also: #3655 |
I tried to evaluate feasibility of adding the error message, but it seems it's probably easier to just fix #3655 than this. The check is done in backend:
UI only gets information about which metrics were successfully logged, it couldn't know which didn't. A proper implementation of the error message should let backend persist errors/warnings related to a run and let UI query that. A hacky workaround is letting UI parse the mlpipeline metric artifact and verify if any metrics are not conformed. However, this cannot scale if backend changes implementation... I'll try if raising the restriction can easily work by just changing the regex. If yes, let's go that route instead and hope fewer people will run into this. |
Further updates will be tracked in #3655 |
@hongye-sun Do you remember the reason for that check? |
@Ark-kun We asked @hongye-sun in another issue report: #3309 (comment). There wasn't any implementation requirement, just resource naming alignment. So I'd prefer we remove this limitation. |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. |
/lifecycle frozen |
What steps did you take:
Actually it is reported by one of our friends from CAIP training team
He was trying to add custom pipeline metrics to one of the pipeline they're authoring and found it painful.
What happened:
The metrics UI did not show up. It turns out that the regex they're using is culprit. However the error message was not very helpful during the efforts.
It would be good to have a better and more informative debug messages.
Environment:
How did you deploy Kubeflow Pipelines (KFP)?
Kubeflow full-fledged deployment.
KFP version:
KFP SDK version:
Anything else you would like to add:
[Miscellaneous information that will assist in solving the issue.]
/kind feature
/area frontend
/priority p2
The text was updated successfully, but these errors were encountered: