INFER ability without N1QL

We are working with partners who are helping customers with Document Modeling, Reverse and Forward engineering with the DataSets.

Some of our largest customers, still use Couchbase as a pure-play KV document store and do not want to or need to use N1QL for some usecases.

Would be useful to have the ability to "sample" the data without a view or index.

I know we have it on the "Documents" tab, its not just an officially supported or documented API

http://127.0.0.1:8091/ui/index.html#/documents?documentsBucket=default&pageLimit=5&pageNumber=0

I am unsure what this will tolerate from a RPS or Volume Standpoint. But would love to sample say 1000 or 2500 documnts off this API, assuming this just doesnt give me X documents from the first vbucket. I assuming this is simply a Key/Document Listing from Memcached