• Provider, endpoint path, model identifier, and SDK or client version.
• Language, runtime, and where the request is executed.
• A minimal request with credentials and private content removed.
• HTTP status, sanitized response body, and request identifier if available.
• Expected response and what actually happened.
• Whether the same request succeeds outside the application.
• Recent changes and steps already attempted.

Replace secrets with an obvious placeholder and check screenshots as well as text. Keep enough structure for another person to understand the request.

When answering, identify which part of the report supports your diagnosis. If the cause is uncertain, propose one focused check and ask the original poster to share its result.

• Question: A shareable example of the user input.
• Expected evidence: The document passage you expected the system to use.
• Retrieved material: The chunks returned, in their original order.
• Configuration: Embedding model, chunking, filters, retrieval count, and any reranking.
• Answer input: The instructions and context actually passed to generation.
• Final answer: The relevant excerpt and the specific error.

Use public documents or a small synthetic example when your corpus is private. Include document versions and a date for the test.

Which stage would you inspect next, and what observation would support changing it? Share results with the original example so the discussion remains reproducible.

A useful planning post includes:

• The task and expected number of requests.
• Estimated input and output tokens per request.
• The unit prices and any pricing categories that apply.
• Expected retries, evaluation runs, and other separately billed services.
• A low, expected, and high usage scenario.

You can organize the token portion with the AI API Cost Calculator. Treat the result as a planning estimate and state which items you included or excluded.

Share an anonymized example of your assumptions. Replies should focus on missing inputs or useful measurements, and include a source for any provider-specific pricing claim.