Hello @RE2L!
Thank you for letting us know, we have some other users who have also been mysteriously getting 139 Error on their workflows. I will let the eng team know about your case and keep you updated.
Chunking is definitely the best practice for large amounts of data and can be very useful for reducing memory load and improving workflow performance.
It sounds like you are using the loop correctly, as the built in loop block is best practice as opposed to using a .map in a JS code block, but I would need more details/screenshots to fully confirm.
Retool enterprise users have been able to work with very large amounts of data so I do not believe it is a fundamental limitation. Hopefully we can get to the bottom of the 139 Error as a bug we can fix quickly.
We can also try to live trouble shoot during office hours!