Question for Pinecone users with multi-tenant applications: When you’re thinking about Pinecone workloads and costs, do you think in terms of indexes or namespaces?
As in:
- A) “I’ll need 2 indexes, each with X vectors and a cost of $Y per index.” OR…
- B) “I’ll need 1,000 namespaces, each with X vectors on average and a cost of $Y per namespace.”