I’ve used the zip file when the zip is a collection of docs, one json object per file. I don’t think cbdocloader supports one giant file with tons of JSON objects in it. That’s what the error is telling you.
I understand that loading a huge number of records at once is not optimal, but I would expect cbdocloader to load say, an array of 1K or 10K object at a time. What is the approach that other people use if the data to be imported is in the millions of documents? Litter them across the file system? This doesn’t seem optimal at all.