ling/hermes-agent
1
0
Fork 0
mirror of https://github.com/NousResearch/hermes-agent.git synced 2026-04-28 06:51:16 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix(logging): extract useful info from HTML error pages, dump debug on max retries

Browse Source 
Three problems with API error debugging:

1. Terminal showed str(error)[:200] — raw HTML gibberish for Cloudflare
   502/503 pages instead of "502 Bad Gateway"
2. errors.log dumped the entire HTML page as unstructured text
3. _dump_api_request_debug was never called when retries exhausted,
   only for non-retryable 4xx errors

Adds _summarize_api_error() that extracts <title> and Cloudflare Ray ID
from HTML error pages, and falls back to SDK error body messages. Now
the terminal shows clean one-liners like:

  📝 Error: HTTP 502 — openrouter.ai | 502: Bad gateway — Ray 9e226...

Also calls _dump_api_request_debug on max_retries_exhausted so the full
request context is written to ~/.hermes/sessions/ for post-mortem.

Made-with: Cursor
This commit is contained in:
Teknium
2026-03-25 18:34:22 -07:00
parent 7258311710
commit 5b29ff50f8
1 changed files with 56 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

64
run_agent.py
View File

@@ -1812,6 +1812,47 @@ class AIAgent:
trajectory = self._convert_to_trajectory_format(messages, user_query, completed)
_save_trajectory_to_file(trajectory, self.model, completed)
@staticmethod
def _summarize_api_error(error: Exception) -> str:
"""Extract a human-readable one-liner from an API error.
Handles Cloudflare HTML error pages (502, 503, etc.) by pulling the
<title> tag instead of dumping raw HTML. Falls back to a truncated
str(error) for everything else.
"""
import re as _re
raw = str(error)
# Cloudflare / proxy HTML pages: grab the <title> for a clean summary
if "<!DOCTYPE" in raw or "<html" in raw:
m = _re.search(r"<title[^>]*>([^<]+)</title>", raw, _re.IGNORECASE)
title = m.group(1).strip() if m else "HTML error page (title not found)"
# Also grab Cloudflare Ray ID if present
ray = _re.search(r"Cloudflare Ray ID:\s*<strong[^>]*>([^<]+)</strong>", raw)
ray_id = ray.group(1).strip() if ray else None
status_code = getattr(error, "status_code", None)
parts = []
if status_code:
parts.append(f"HTTP {status_code}")
parts.append(title)
if ray_id:
parts.append(f"Ray {ray_id}")
return " — ".join(parts)
# JSON body errors from OpenAI/Anthropic SDKs
body = getattr(error, "body", None)
if isinstance(body, dict):
msg = body.get("error", {}).get("message") if isinstance(body.get("error"), dict) else body.get("message")
if msg:
status_code = getattr(error, "status_code", None)
prefix = f"HTTP {status_code}: " if status_code else ""
return f"{prefix}{msg[:300]}"
# Fallback: truncate the raw string but give more room than 200 chars
status_code = getattr(error, "status_code", None)
prefix = f"HTTP {status_code}: " if status_code else ""
return f"{prefix}{raw[:500]}"
def _mask_api_key_for_logs(self, key: Optional[str]) -> Optional[str]:
if not key:
return None
@@ -6363,16 +6404,16 @@ class AIAgent:
retry_count += 1
elapsed_time = time.time() - api_start_time
# Enhanced error logging
error_type = type(api_error).__name__
error_msg = str(api_error).lower()
_error_summary = self._summarize_api_error(api_error)
logger.warning(
"API call failed (attempt %s/%s) error_type=%s %s error=%s",
"API call failed (attempt %s/%s) error_type=%s %s summary=%s",
retry_count,
max_retries,
error_type,
self._client_log_context(),
api_error,
_error_summary,
)
_provider = getattr(self, "provider", "unknown")
@@ -6382,9 +6423,8 @@ class AIAgent:
self._vprint(f"{self.log_prefix}⚠️ API call failed (attempt {retry_count}/{max_retries}): {error_type}{_status_code_str}", force=True)
self._vprint(f"{self.log_prefix} 🔌 Provider: {_provider} Model: {_model}", force=True)
self._vprint(f"{self.log_prefix} 🌐 Endpoint: {_base}", force=True)
cleaned_error = self._clean_error_message(str(api_error))
self._vprint(f"{self.log_prefix} 📝 Error: {cleaned_error}", force=True)
if status_code == 400:
self._vprint(f"{self.log_prefix} 📝 Error: {_error_summary}", force=True)
if status_code and status_code < 500:
_err_body = getattr(api_error, "body", None)
_err_body_str = str(_err_body)[:300] if _err_body else None
if _err_body_str:
@@ -6640,9 +6680,17 @@ class AIAgent:
if self._try_activate_fallback():
retry_count = 0
continue
_final_summary = self._summarize_api_error(api_error)
self._vprint(f"{self.log_prefix}❌ Max retries ({max_retries}) exceeded. Giving up.", force=True)
logging.error(f"{self.log_prefix}API call failed after {max_retries} retries. Last error: {api_error}")
logging.error(f"{self.log_prefix}Request details - Messages: {len(api_messages)}, Approx tokens: {approx_tokens:,}")
self._vprint(f"{self.log_prefix} 💀 Final error: {_final_summary}", force=True)
logging.error(
"%sAPI call failed after %s retries. %s | provider=%s model=%s msgs=%s tokens=~%s",
self.log_prefix, max_retries, _final_summary,
_provider, _model, len(api_messages), f"{approx_tokens:,}",
)
self._dump_api_request_debug(
api_kwargs, reason="max_retries_exhausted", error=api_error,
)
raise api_error
wait_time = min(2 ** retry_count, 60) # Exponential backoff: 2s, 4s, 8s, 16s, 32s, 60s, 60s