mirror of
https://github.com/open-webui/open-webui.git
synced 2026-11-09 10:50:35 -06:00
[GH-ISSUE #4884] Any chat that takes longer than 30 seconds to reply fails, very long context chats always fail first time #52439
Reference in New Issue
Block a user
Originally created by @matatonic on GitHub (Aug 24, 2024).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/4884
Bug Report
Installation Method
Official docker
Environment
Open WebUI Version: v0.3.14
Ollama (if applicable): NA
Operating System: Ubuntu 22.04
Browser (if applicable): Firefox 129.0.2 or Chromium 127.0.6533.99
Confirmation:
Expected Behavior:
A long context chat, or slow operation (such as generating many minutes of speech audio) should continue to work after more than 30 seconds.
Actual Behavior:
In a long context chat, starting from a cold context cache, a chat request can take a while to warm up the cache, but after 30 seconds without any reply the UI seems to fail and no text is produced. With very large context (100k) even api.groq.com isn't fast enough to start producing text in this time.
Description
Bug Summary:
Any chat (or related action, such as generating speech) which takes more than 30 seconds is aborted.
Reproduction Details
Steps to Reproduce:
Load a 100k context into chat, ask for a reply.
Logs and Screenshots
Browser Console Logs:
N/A
Docker Container Logs:
Screenshots/Screen Recordings (if applicable):
Additional Information
My whole setup uses various openai compatible API's, so I don't know if it's related to anything else, or if it's an openai api client issue. The default OpenAI API client timeout should be 10 minutes. I've tried setting:
But that doesn't fix anything, still failed after 30 seconds.