[BUG] UI crashes or is extremely slow from too many columns

[adam-postera]

Willingness to contribute

No. I cannot contribute a bug fix at this time.

MLflow version

2.8.1

System information

Running on AWS ECS and exposing with a load balancer. Python3.10

Describe the problem

Our models have a lot of metrics that we are logging and this is dramatically slowing down the UI to the point where it takes very long to load or crashes completely. We would love if we would have the option to not preload the columns in the UI and only load the ones after they are selected. We ran the DevTools and saw it is the search is taking a long time, with the majority of the time waiting for the server. We also ran performance and saw these issues, but this was after we stopped it loading after one minute of time.

Steps to reproduce the bug

Load mlflow experiment with 12000 columns

Code to generate data required to reproduce the bug

No response

Is the console panel in DevTools showing errors relevant to the bug?

No response

Does the network panel in DevTools contain failed requests relevant to the bug?

No response

[github-actions]

@mlflow/mlflow-team Please assign a maintainer and start triaging this issue.