Emre Yilmaz

For those running at that scale: is it the machine, the proxies, or the process that becomes the constraint?

A summary line at the end of an import would fix this: rows read, profiles created, rows skipped and why. Right now the count difference is the only signal and you have to go looking for the cause. Until then, sort the CSV by the column you consider unique before importing — duplicates become obvious in the file, which is a better place to catch them than after the fact.

Had exactly this last week. Restarting the profile rather than the whole app sorted it.

This matches my experience. Consistency has mattered more than uniqueness for me.