[GH-ISSUE #623] Chat assistant error does not appear in conversation if more than 30 seconds pass without receiving response from pipe function

bsakaguchi commented

2026-10-18 14:08:53 +00:00

Owner

Originally created by @QxNam on GitHub (Jul 14, 2025).
Original GitHub issue: https://github.com/open-webui/docs/issues/623

I use open-webui version v0.6.10.
I have tested many chat cases using n8n-pipe function and only chats with response < 30 seconds show chat assistant, the rest do not show (reload web shows). Why? How to improve response time?
This is the template:

"""
title: n8n Pipe Function
author: Cole Medin
author_url: https://www.youtube.com/@ColeMedin
version: 0.2.0

This module defines a Pipe class that utilizes N8N for an Agent
"""

from typing import Optional, Callable, Awaitable
from pydantic import BaseModel, Field
import os
import time
import requests


def extract_event_info(event_emitter) -> tuple[Optional[str], Optional[str]]:
    if not event_emitter or not event_emitter.__closure__:
        return None, None
    for cell in event_emitter.__closure__:
        if isinstance(request_info := cell.cell_contents, dict):
            chat_id = request_info.get("chat_id")
            message_id = request_info.get("message_id")
            return chat_id, message_id
    return None, None


class Pipe:
    class Valves(BaseModel):
        n8n_url: str = Field(
            default="https://n8n.[your domain].com/webhook/[your webhook URL]"
        )
        n8n_bearer_token: str = Field(default="...")
        input_field: str = Field(default="chatInput")
        response_field: str = Field(default="output")
        emit_interval: float = Field(
            default=2.0, description="Interval in seconds between status emissions"
        )
        enable_status_indicator: bool = Field(
            default=True, description="Enable or disable status indicator emissions"
        )

    def __init__(self):
        self.type = "pipe"
        self.id = "n8n_pipe"
        self.name = "N8N Pipe"
        self.valves = self.Valves()
        self.last_emit_time = 0
        pass

    async def emit_status(
        self,
        __event_emitter__: Callable[[dict], Awaitable[None]],
        level: str,
        message: str,
        done: bool,
    ):
        current_time = time.time()
        if (
            __event_emitter__
            and self.valves.enable_status_indicator
            and (
                current_time - self.last_emit_time >= self.valves.emit_interval or done
            )
        ):
            await __event_emitter__(
                {
                    "type": "status",
                    "data": {
                        "status": "complete" if done else "in_progress",
                        "level": level,
                        "description": message,
                        "done": done,
                    },
                }
            )
            self.last_emit_time = current_time

    async def pipe(
        self,
        body: dict,
        __user__: Optional[dict] = None,
        __event_emitter__: Callable[[dict], Awaitable[None]] = None,
        __event_call__: Callable[[dict], Awaitable[dict]] = None,
    ) -> Optional[dict]:
        await self.emit_status(
            __event_emitter__, "info", "Đang nhập tin nhắn...", False
        )
        chat_id, _ = extract_event_info(__event_emitter__)
        messages = body.get("messages", [])

        # Verify a message is available
        if messages:
            question = messages[-1]["content"]
            try:
                # Invoke N8N workflow
                headers = {
                    "Authorization": f"Bearer {self.valves.n8n_bearer_token}",
                    "Content-Type": "application/json",
                }
                payload = {"sessionId": f"{chat_id}"}
                payload[self.valves.input_field] = question
                payload["user"] = __user__
                response = requests.post(
                    self.valves.n8n_url, json=payload, headers=headers
                )
                if response.status_code == 200:
                    n8n_response = response.json()[self.valves.response_field]
                else:
                    raise Exception(f"Error: {response.status_code} - {response.text}")

                # Set assitant message with chain reply
                body["messages"].append({"role": "assistant", "content": n8n_response})
            except Exception as e:
                await self.emit_status(
                    __event_emitter__,
                    "error",
                    f"Error during sequence execution: {str(e)}",
                    True,
                )
                return {"error": str(e)}
        # If no message is available alert user
        else:
            await self.emit_status(
                __event_emitter__,
                "error",
                "No messages found in the request body",
                True,
            )
            body["messages"].append(
                {
                    "role": "assistant",
                    "content": "No messages found in the request body",
                }
            )

        await self.emit_status(__event_emitter__, "info", "", True)
        return n8n_response

Response from n8n-pipe after 30s will stay like this forever, only when reloading the web will it display.

Originally created by @QxNam on GitHub (Jul 14, 2025). Original GitHub issue: https://github.com/open-webui/docs/issues/623 I use open-webui version `v0.6.10`. I have tested many chat cases using n8n-pipe function and only chats with response < 30 seconds show chat assistant, the rest do not show (reload web shows). Why? How to improve response time? This is the template: ``` """ title: n8n Pipe Function author: Cole Medin author_url: https://www.youtube.com/@ColeMedin version: 0.2.0 This module defines a Pipe class that utilizes N8N for an Agent """ from typing import Optional, Callable, Awaitable from pydantic import BaseModel, Field import os import time import requests def extract_event_info(event_emitter) -> tuple[Optional[str], Optional[str]]: if not event_emitter or not event_emitter.__closure__: return None, None for cell in event_emitter.__closure__: if isinstance(request_info := cell.cell_contents, dict): chat_id = request_info.get("chat_id") message_id = request_info.get("message_id") return chat_id, message_id return None, None class Pipe: class Valves(BaseModel): n8n_url: str = Field( default="https://n8n.[your domain].com/webhook/[your webhook URL]" ) n8n_bearer_token: str = Field(default="...") input_field: str = Field(default="chatInput") response_field: str = Field(default="output") emit_interval: float = Field( default=2.0, description="Interval in seconds between status emissions" ) enable_status_indicator: bool = Field( default=True, description="Enable or disable status indicator emissions" ) def __init__(self): self.type = "pipe" self.id = "n8n_pipe" self.name = "N8N Pipe" self.valves = self.Valves() self.last_emit_time = 0 pass async def emit_status( self, __event_emitter__: Callable[[dict], Awaitable[None]], level: str, message: str, done: bool, ): current_time = time.time() if ( __event_emitter__ and self.valves.enable_status_indicator and ( current_time - self.last_emit_time >= self.valves.emit_interval or done ) ): await __event_emitter__( { "type": "status", "data": { "status": "complete" if done else "in_progress", "level": level, "description": message, "done": done, }, } ) self.last_emit_time = current_time async def pipe( self, body: dict, __user__: Optional[dict] = None, __event_emitter__: Callable[[dict], Awaitable[None]] = None, __event_call__: Callable[[dict], Awaitable[dict]] = None, ) -> Optional[dict]: await self.emit_status( __event_emitter__, "info", "Đang nhập tin nhắn...", False ) chat_id, _ = extract_event_info(__event_emitter__) messages = body.get("messages", []) # Verify a message is available if messages: question = messages[-1]["content"] try: # Invoke N8N workflow headers = { "Authorization": f"Bearer {self.valves.n8n_bearer_token}", "Content-Type": "application/json", } payload = {"sessionId": f"{chat_id}"} payload[self.valves.input_field] = question payload["user"] = __user__ response = requests.post( self.valves.n8n_url, json=payload, headers=headers ) if response.status_code == 200: n8n_response = response.json()[self.valves.response_field] else: raise Exception(f"Error: {response.status_code} - {response.text}") # Set assitant message with chain reply body["messages"].append({"role": "assistant", "content": n8n_response}) except Exception as e: await self.emit_status( __event_emitter__, "error", f"Error during sequence execution: {str(e)}", True, ) return {"error": str(e)} # If no message is available alert user else: await self.emit_status( __event_emitter__, "error", "No messages found in the request body", True, ) body["messages"].append( { "role": "assistant", "content": "No messages found in the request body", } ) await self.emit_status(__event_emitter__, "info", "", True) return n8n_response ``` Response from n8n-pipe after 30s will stay like this forever, only when reloading the web will it display. <img width="348" height="61" alt="Image" src="https://github.com/user-attachments/assets/1df7c331-5ac6-4766-9ec6-e8f37d49ab4d" />

bsakaguchi closed this issue

2026-10-18 14:08:53 +00:00

bsakaguchi commented

2026-10-18 14:08:54 +00:00

Author

Owner

@Classic298 commented on GitHub (Jul 14, 2025):

Check your http timeout env vars and check your http web proxy. 100% your error is somewhere there with wrong configuration of connection handling and timeouts

@Classic298 commented on GitHub (Jul 14, 2025): Check your http timeout env vars and check your http web proxy. 100% your error is somewhere there with wrong configuration of connection handling and timeouts

bsakaguchi commented

2026-10-18 14:08:54 +00:00

Author

Owner

@QxNam commented on GitHub (Jul 14, 2025):

Check your http timeout env vars and check your http web proxy. 100% your error is somewhere there with wrong configuration of connection handling and timeouts

I using nginx and configured:

proxy_connect_timeout 300s;
proxy_send_timeout 300s;
proxy_read_timeout 300s;

but still can't change anything!!!

@QxNam commented on GitHub (Jul 14, 2025): > Check your http timeout env vars and check your http web proxy. 100% your error is somewhere there with wrong configuration of connection handling and timeouts I using nginx and configured: ``` proxy_connect_timeout 300s; proxy_send_timeout 300s; proxy_read_timeout 300s; ``` but still can't change anything!!!

bsakaguchi commented

2026-10-18 14:08:54 +00:00

Author

Owner

@Classic298 commented on GitHub (Sep 14, 2025):

@QxNam sorry for the late followup - please increase the AIOHTTP timeout env var

@Classic298 commented on GitHub (Sep 14, 2025): @QxNam sorry for the late followup - please increase the AIOHTTP timeout env var

bsakaguchi referenced this issue

2026-10-18 14:10:26 +00:00

[PR #88] [MERGED] update web search docs for 0.2.0 release #238

Rows
Columns

[GH-ISSUE #623] Chat assistant error does not appear in conversation if more than 30 seconds pass without receiving response from pipe function #88