[FEA]: Fully in-process Service query

Is this a new feature, an improvement, or a change to existing functionality?

Improvement

How would you describe the priority of this feature request

Significant improvement

Please provide a clear description of problem this feature solves

We should be able to run service without an external VDB sidecar when using retriever service mode

Describe the feature, and optionally a solution or implementation and any alternatives

Retriever Service currently requires a separate VDB sidecar for queries; disabling VDB causes query requests to fail with a 404. This complicates ingest-and-query workflows and may require running two services, and potentially loading a second embedding model.

The requested feature is to make VDB support part of the Retriever Service startup flow, ideally through a single command, while preserving compatibility with the existing Helm deployment architecture.

Additional context

No response

Start by tracing Retriever Service startup and query handling with VDB disabled, then review the existing Helm deployment architecture. Confirm how VDB should initialize through a single command while preserving the sidecar-compatible path. Done means retriever queries work without an external VDB sidecar and existing Helm deployments remain compatible.