## Problem Current batch ingestion provides basic success/failure reporting. However: - Errors are not structured or categorized - Users cannot easily identify which items failed and why - No support for retry strategies ## Goal Improve error handling and reporting for batch ingestion. ## Proposed Improvements - Return structured response: - success_count - failure_count - successful items - failed items with error messages - Classify errors: - Timeout - Validation errors - LLM failures - Internal errors - Optional future: - Retry failed items - Partial retry endpoint ## Example Response ```json { "success_count": 48, "failure_count": 2, "results": [...], "errors": [ {"index": 3, "error": "LLM timeout"}, {"index": 7, "error": "Invalid input"} ] } #143
Problem
Current batch ingestion provides basic success/failure reporting. However:
Goal
Improve error handling and reporting for batch ingestion.
Proposed Improvements
Return structured response:
Classify errors:
Optional future:
Example Response
{ "success_count": 48, "failure_count": 2, "results": [...], "errors": [ {"index": 3, "error": "LLM timeout"}, {"index": 7, "error": "Invalid input"} ] } #143