The Query documentation is woefully thin on the subject of query context, the only meaningful description that I could find being
this and this.
We should have documentation on the concept of query context and how it is used as a connection string, so that Query requests don't have to have absolute paths to collections, and instead can use relative paths, which are resolved via the use of the query_context.
We should have a whole section on tenant separation and how this is implemented using the query context: the same application can be switched from using buckets or collections, or can be switched from using one scope or another by just pointing the query context to the required scope.
The prepared statements, PREPARE , EXECUTE and system:prepareds sections in particular should be expanded to explain that prepared statements are relative to the query_context used, so that it is possible to create multiple prepared statements with the same name, each referencing different query_contexts.
This should be urgently documented in current versions, as the implications of not having it documented are that it gives the impression that Query cannot handle multi-tenancy, which couldn't be farther from the truth.