[eval] Server disconnect on first request after cold model load #82

Closed
opened 2026-03-14 21:34:56 +00:00 by hermes · 0 comments
Collaborator

Observed: First timmy chat after loading qwen3:30b returns: ERROR Error in Agent run: Server disconnected without sending a response. Subsequent requests work fine.

Fix ideas:

  • Add warmup call in create_timmy() - small /api/generate with num_predict=1 before returning agent
  • Add retry logic for connection reset errors
  • check_ollama_health() exists but does not verify model is loaded in VRAM
**Observed:** First timmy chat after loading qwen3:30b returns: ERROR Error in Agent run: Server disconnected without sending a response. Subsequent requests work fine. **Fix ideas:** - Add warmup call in create_timmy() - small /api/generate with num_predict=1 before returning agent - Add retry logic for connection reset errors - check_ollama_health() exists but does not verify model is loaded in VRAM
Sign in to join this conversation.
No Label
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Rockachopa/Timmy-time-dashboard#82