srbhr/Resume-Matcher · error · ValueError

Failed after {retries + 1} attempts

Error message

Failed after {retries + 1} attempts

What it means

complete_json loops for up to retries + 1 attempts; if the loop exits without returning (all attempts consumed by content-quality failures that didn't raise), it raises this generic ValueError. Transport errors are re-raised as-is since the LiteLLM Router already retried with backoff.

Source

Thrown at apps/backend/app/llm.py:1360

            ):
                json_mode_failed = True
                logging.warning(
                    "Provider rejected response_format for %s; falling back to "
                    "prompt-only JSON mode (attempt %d)",
                    model_name,
                    attempt + 1,
                )
                if attempt < retries:
                    continue
            raise

        except Exception:
            # Transport errors — Router already retried with backoff.
            # Cooldowns are disabled (see _build_router); no additional
            # retry is attempted here.
            raise

    raise ValueError(f"Failed after {retries + 1} attempts")

View on GitHub (pinned to 116f9cc3b0)

Solutions

1. Increase retries in the complete_json call
2. Check server logs for the per-attempt warnings ('Content extraction failed (attempt N)') to find the root cause
3. Switch to a stronger model or enable JSON mode for the request
4. Add a fallback UX (cached result or graceful degradation) in the calling feature (analyze_resume, generate_enhancements, etc.)

async def llm_healthy() -> bool:
    try:
        await complete_json('Return {"ok": true}', retries=1)
        return True
    except ValueError:
        return False

try:
    data = await complete_json(prompt)
except ValueError:
    # includes 'Failed after N attempts'
    return serve_cached_or_default_result()
except Exception:
    # transport errors — Router already retried; surface as 503
    raise ServiceUnavailable("LLM provider unreachable")

When it happens

Trigger: Every attempt of complete_json fails with empty responses or extraction failures, the retry loop is exhausted, and the final raise at the end of the function is reached.

Common situations: Sustained provider degradation returning empty/bad content across all retries; retries configured to 0; model systematically unable to satisfy the prompt's JSON requirement.

Related errors

• Failed to parse JSON after {retries + 1} attempts: {e}
• LLM completion failed. Please check your API configuration a
• JSON extraction exceeded max recursion depth: {_depth}
• Content too large for JSON extraction: {len(content)} bytes
• No JSON found in response: {original[:200]}

AI-assisted analysis of srbhr/Resume-Matcher@116f9cc3b0 (2026-08-28). Data as JSON: /api/errors/1c31033541fd5efc. Report an issue: GitHub.