seb/hermes-agent
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(gateway): recover model on post-interrupt turn; gate fallback status (#35381)

Browse Source 
Empty model could reach the API on a recovery turn after stream_interrupt_abort,
failing HTTP 400 "No models provided" with no recovery — the session went
silent until the user manually re-sent (#35314).

- gateway/run.py: cache last-successfully-resolved model per session (+ a
  process-wide slot); when a fresh config read returns an empty model on a
  recovery turn, reuse the last-known-good instead of building model="".
- run_agent.py + agent/conversation_loop.py: only emit "trying fallback..."
  status when a fallback chain actually exists, so the UI stops announcing a
  fallback that will never run (also #17446).
- tests: empty-model recovery + _has_pending_fallback gate.
This commit is contained in:
Teknium
2026-05-30 07:28:06 -07:00
committed by GitHub
parent 10dec7c6dc
commit 2b16b756a7
4 changed files with 208 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files
agent/conversation_loop.py
+15 -8
View File
@@ -1492,7 +1492,8 @@ def run_conversation(
if retry_count >= max_retries:
# Try fallback before giving up
agent._buffer_status(f"⚠️ Max retries ({max_retries}) for invalid responses — trying fallback...")
if agent._has_pending_fallback():
agent._buffer_status(f"⚠️ Max retries ({max_retries}) for invalid responses — trying fallback...")
if agent._try_activate_fallback():
retry_count = 0
compression_attempts = 0
@@ -3094,12 +3095,17 @@ def run_conversation(
) and not is_context_length_error
if is_client_error:
# Try fallback before aborting — a different provider
# may not have the same issue (rate limit, auth, etc.)
if classified.reason == FailoverReason.content_policy_blocked:
agent._buffer_status("⚠️ Provider safety filter blocked this request — trying fallback...")
else:
agent._buffer_status(f"⚠️ Non-retryable error (HTTP {status_code}) — trying fallback...")
# Try fallback before aborting — a different provider may
# not have the same issue (rate limit, auth, etc.). Only
# announce the attempt when a fallback chain actually
# exists; otherwise "trying fallback..." is a lie and the
# session looks like it's recovering when it's about to
# abort silently (#35314, #17446).
if agent._has_pending_fallback():
if classified.reason == FailoverReason.content_policy_blocked:
agent._buffer_status("⚠️ Provider safety filter blocked this request — trying fallback...")
else:
agent._buffer_status(f"⚠️ Non-retryable error (HTTP {status_code}) — trying fallback...")
if agent._try_activate_fallback():
retry_count = 0
compression_attempts = 0
@@ -3242,7 +3248,8 @@ def run_conversation(
retry_count = 0
continue
# Try fallback before giving up entirely
agent._buffer_status(f"⚠️ Max retries ({max_retries}) exhausted — trying fallback...")
if agent._has_pending_fallback():
agent._buffer_status(f"⚠️ Max retries ({max_retries}) exhausted — trying fallback...")
if agent._try_activate_fallback():
retry_count = 0
compression_attempts = 0