Currently users do not receive information when the memory usage of the computing environment is maxed out. When the notebook starts to train the model, it will run for a few seconds up to a few minutes, and then it seems to just hang.the notebook execution seems to just hang and never complete, even after a couple of weeks since the notebooks have been sitting there and the kernel connection is never established again. The output never seems to update.
It would have been helpful if Watson Studio gave users some indication that memory usage was maxed out, or another performance indicator, so users would have a hint that I need to get a bigger sized environment. That would help the customer avoid some frustrations, increase the size of their environment, and then be successful with Watson Studio.
Why is it useful?
|Who would benefit from this IDEA?||As a data scientist uses jupyter notebook, I'd like to get informed when the computer enviroment that the notebook is running on is not enough for the model.|
How should it work?
|Submitting Organization||Offering Management|