Couchbase's primary index is a sorted list of all document keys within a bucket, maintained asynchronously and supporting various scan consistencies. It allows full bucket scans when queries don't have filters or other indexes can be used, but this comes at a high cost due to the need to fetch and filter all documents. The primary index should not be used in production unless necessary, as it's slow and unnecessary document fetches waste memory and CPU resources. However, it can be useful for development, querying specific ranges of keys, or when the full document key is known. Named primary indexes with multiple replicas offer high availability and query load distribution benefits. The Index Advisor in Couchbase 6.5 analyzes N1QL statements or workloads to advise suitable secondary indexes but not primary indexes, as they're not recommended due to performance issues. When creating multiple "named" indexes of the same type and combination, it's essential to consider the cost tradeoff, as maintaining multiple index nodes can impact performance. There is no other way to retrieve a document ID without using indexes or views if the ID is unknown in advance, but you can use the `USE KEYS` clause to fetch documents directly without consulting the primary index.