Unexplained difference between the elapsed time client side and the metrics

[drigby]

@vsr1

2. As your doing limit 1 and if you already know the document id use USE KEYS to avoid index scan this same as bucket.Get using N1QL.

Or if you literally just want a single key and the catalog element in the JSON doc then just use K/V Sub-document API (SUBDOC_GET) - that’ll always be faster.