Namespaces And Multitenant Isolation
Pinecone · Pinecone
Vector Database — Pinecone
Pinecone evals — Namespaces & Multi-tenant Isolation (relift v3 InfraRed)
About Pinecone
Pinecone is a managed vector database for AI applications — serverless and pod-based indexes, namespaces for multi-tenant isolation, hybrid sparse-dense search, integrated inference (embed + rerank), and Pinecone Assistant for retrieval-augmented generation with citations.
Sample tests· showing 3 of 9
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | Operator stores 500 tenants' vectors in one default namespace and filters by metadata.tenant_id at query time. | Per docs, namespace is the recommended multi-tenant primitive. Use one namespace per tenant — query/upsert/delete are scoped to namespace. Metadata-filter-only isolation leaks under filter bugs and wastes ANN work on out-of-tenant candidates. Migrate to namespace-per-tenant. | Pass / FailAi Platformcritical |
| 02 | Operator needs to audit which tenants currently have vectors in the index. | Call describe_index_stats — response includes per-namespace vector counts. Use this for audit, billing attribution, and lifecycle (purge empty namespaces). Do not rely on operator-side tenant registry alone — verify against Pinecone state. | Pass / FailAi Platformmedium |
| 03 | Operator wants to query across 5 namespaces in one call. | Per docs, query is single-namespace per call. To search across N namespaces, issue N parallel queries and merge client-side (e.g. RRF or score normalization). Do not try to pass a list of namespaces — not supported. | Pass / FailAi Platformhigh |
How this eval is graded
Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3.
Rubric criteria
- Pinecone
- Ai Platform
- Namespaces And Multitenant Isolation
Recommended for
Works with
Related evals
Run this eval in your workspace
Connect your data, configure thresholds, and review results with your team.