Understanding RagManagedDb

This page explains RagManagedDb, its underlying technology, its role in Vertex AI RAG Engine, and how to manage its configuration.

This page covers the following topics:

What is RagManagedDb?

RagManagedDb is an enterprise-ready, fully-managed Google Spanner instance that Vertex AI RAG Engine uses to store its resources. You can also optionally use it as the vector database of choice for your RAG corpora.

By using Spanner, Vertex AI RAG Engine provides a consistent, highly available, and scalable database for your application. To learn more about Google Spanner, see Spanner.

Vertex AI RAG Engine always stores your RAG corpus and file resource metadata in RagManagedDb, even if you use a different vector database. When you use RagManagedDb as your vector database, it also stores and manages the vector embeddings of your documents. The vector database retrieves relevant documents based on the semantic similarity between a document and a given query.

Manage tiers

You can scale your RagManagedDb instance based on your usage and performance requirements by selecting a tier. The tier is a project-level setting that you configure in the RagEngineConfig resource.

The following tiers are available:

  • Scaled tier: A production-scale instance recommended for large datasets, performance-sensitive workloads, and production environments. This tier configures the underlying Spanner instance to autoscale from 1 to 10 nodes (1,000 to 10,000 processing units).

  • Basic tier (Default): A cost-effective, low-compute instance fixed at 100 processing units (0.1 nodes). This tier is recommended for the following use cases:

    • Experimenting with RagManagedDb.
    • Small data sizes.
    • Latency-insensitive workloads.
    • Using Vertex AI RAG Engine with other vector databases.

  • Unprovisioned tier: Deletes the RagManagedDb instance and all associated data, which stops service billing. This action is irreversible and disables the Vertex AI RAG Engine service for the project. To use the service again, you must update the tier by calling the UpdateRagEngineConfig API. For more information on billing, see Vertex AI RAG Engine billing.

Get the project configuration

To retrieve your current project configuration, use the GetRagEngineConfig API. For code samples, see the following:

Update the project configuration

To change your project's tier, use the UpdateRagEngineConfig API. For code samples, see the following:

What's next