Managed spanner in Vertex AI RAG Engine

Google Cloud Database
1

Hi everyone,

I’m currently using Vertex AI RAG Engine with the default, managed Spanner (RagManagedDb), but I’m encountering issues with monitoring and managing it. Does anyone know how to address these concerns?

  1. Spanner Costs (which worries me): I’m not seeing any RagManagedDb costs in my Google Cloud billing reports. Even with initial free credits, the cost should ideally appear in the billing with a 0.00 amount, indicating service usage. Are these costs “hidden” within other Vertex AI line items, and if so, how can I extract them?

  2. Storage Size: How can I check how much space my corpora and indexes are occupying in the managed Spanner? The lack of this information makes capacity planning difficult.

  3. Spanner Load Metrics: Is there a way to access performance metrics (e.g., CPU utilization, I/O) for Spanner during file uploads or data retrieval in the RAG Engine? I need this for diagnosis and optimization.

I understand that Vertex AI RAG Engine doesn’t physically create a Spanner instance directly assigned to my account, and that Google manages it. Nevertheless, I believe that some basic information about this instance should be available to the user.

I’d be very grateful for any tips or shared experiences!

Thanks!