Database
Pinecone development
How does AonHive use Pinecone?
Pinecone development for a managed vector database and production retrieval.
Pinecone fits when the vector index should be hosted and scaled apart from the application database. The operational row stays in Postgres. The index stores an id and the metadata we filter on. A deleted record is deleted from the index in the same change.Discuss this stackHow we use it
A hosted index, with the record elsewhere.
- 01
Filter before you answer
Tenant, owner, and status are metadata. A query without them can return another customer's text.
- 02
Write the index with the record
An insert, an update, and a delete touch the source and the index. A nightly rebuild is a repair, not the normal path.
- 03
Treat a miss as a state
An empty result, a timeout, or a low score has a response. The product does not invent a paragraph when retrieval failed.
Pinecone
Bring the system this has to fit.
Tell us the product, the estate you already run, and what your team needs to own. We will say whether this belongs in the stack.