admin-valentin/open-notebook
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fix: increase timeout for Ollama and local LLM operations

Browse source 
Users with Ollama reported timeout errors on notebook chat while the
backend was still processing. The answer would appear after refresh.

- Frontend axios timeout: 5 min → 10 min
- Backend chat service timeout: 2 min → 10 min

Local LLMs can take several minutes for complex questions with large
contexts, especially on slower hardware.
This commit is contained in:
LUIS NOVO 2025-12-14 11:31:03 -03:00
parent e5eeb90341
commit 3001537aa7
2 changed files with 6 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
api/chat_service.py
View file

@ -135,7 +135,7 @@ class ChatService:
if model_override is not None: if model_override is not None:
data["model_override"] = model_override data["model_override"] = model_override
async with httpx.AsyncClient(timeout=120.0) as client: # Longer timeout for chat async with httpx.AsyncClient(timeout=600.0) as client: # 10 min timeout for Ollama/local LLMs
response = await client.post( response = await client.post(
f"{self.base_url}/api/chat/execute", f"{self.base_url}/api/chat/execute",
json=data, json=data,

10
frontend/src/lib/api/client.ts
View file

@ -3,12 +3,12 @@ import { getApiUrl } from '@/lib/config'
// API client with runtime-configurable base URL // API client with runtime-configurable base URL
// The base URL is fetched from the API config endpoint on first request // The base URL is fetched from the API config endpoint on first request
// Timeout increased to 5 minutes (300000ms = 300s) to accommodate slow LLM operations // Timeout increased to 10 minutes (600000ms = 600s) to accommodate slow LLM operations
// (transformations, insights generation) especially on slower hardware (Ollama, LM Studio) // (transformations, insights generation, chat) especially on slower hardware (Ollama, LM Studio)
// Note: Frontend uses milliseconds (300000ms), backend uses seconds (300s) - both equal 5 minutes // Note: Frontend uses milliseconds, backend uses seconds
// To configure: Set API_CLIENT_TIMEOUT=600 in .env for 10 minutes (600s = 600000ms) // Local LLMs can take several minutes for complex questions with large contexts
export const apiClient = axios.create({ export const apiClient = axios.create({
timeout: 300000, // 300 seconds = 5 minutes timeout: 600000, // 600 seconds = 10 minutes
headers: { headers: {
'Content-Type': 'application/json', 'Content-Type': 'application/json',
}, },