Pipelines: Pipeline Run UI does not load on large data / 1k pods while executing, when done, displays "No graph to show"

[yoni-taranis]

What happened:
While running a pipeline that contains about 1k pods performing tasks parallelic, the UI does not load after re-loading it from the 1st time entering to the page. It is simply showing the "lodader" icon without stopping.

When the Run finishes, going into it will display "No graph to show" although the pods still exist on the cluster with the Succeeded status.

What did you expect to happen:
Expected to view all the pods and their status.

What steps did you take:
It does not seem to have any backend side errors when looking at the Network section in Dev Tools.

Anything else you would like to add:
We would really like to get this solved and if we can help with the process of doing it we will be glad! Any pointers would be helpful.

I do understand it might be worthless to see a graph of 1k pods (or even less) - but it would actually be helpful to visualize errors (or just pods names with errors) - especially in production env

[Bobgy]

You can follow instructions in https://github.com/kubeflow/pipelines/blob/master/frontend/README.md for running a local UI instance and connect API requests to your real backend.

Feel free to let me know if you have any further questions.

[yoni-taranis]

I will try and take a look.

Just a hunch - I have a feeling it is related to the refresh interval of the UI that is created when entering that screen.

[yoni-taranis]

Just updating, that I've added a few short lines of code to add a "Summary" tab to the view
By not defaulting the graph view (which is very heavy to process I guess on a large amount of pods + retries) we are now utilizing the frontend much more.

[Bobgy]

@yoni-taranis Awesome, that sounds a reasonable solution.

Do you want to contribute the solution here? I think it makes sense if UI detects how many nodes there are in the graph, and if it is huge, fallback to display the overview.

[yoni-taranis]

@Bobgy I will need to refactor it a little bit in order to contribute, I will find the time to do it soon and create a PR.

[stale[bot]]

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.

[stale[bot]]

This issue has been automatically closed because it has not had recent activity. Please comment "/reopen" to reopen it.