Content portal → retrieval service indexing integration

[iliia-veselov]

What is the problem this feature will solve?

Documents reach the retrieval service through an upstream content portal. The API surface that integration needs is being built — #18 update document metadata, #19 indexing status, #20 channel cleanup — but no issue names the integration itself, so nobody owns the end-to-end flow or its acceptance criteria, and there is no single place to see whether it works.

What is the proposed feature or solution?

1. Define the contract between the portal and the retrieval service: publish, update, delete, reindex, read status.
2. Wire the portal to the endpoints from Add endpoint to update document's metadata #18, Add indexing status endpoint #19 and Add endpoint for channel cleanup #20.
3. Handle partial and failed indexing visibly — the portal must be able to show a document as failed rather than silently missing.
4. Verify the flow end to end against a realistic corpus, not test fixtures.

Acceptance criteria

• A document added in the portal becomes searchable without manual steps.
• Removing or replacing a document in the portal is reflected in the index.
• Indexing status, including failures, is visible from the portal.
• The flow is verified against a realistic corpus.