[BUG]WebSim state broadcasting fails inside Docker containers due to an event loop conflict, preventing the frontend from receiving real-time status updates

[Gkkkk]

🐛 Bug: WebSim WebSocket not broadcasting state in Docker (event loop already running)

Environment

• OS: Windows 11 (likely affects other platforms using Docker)
• Docker: Docker Desktop 29.1.3
• Python (container): 3.10.19
• OM1 version: main branch (2026-01-18)

---

Steps to Reproduce

1. Start OM1 with Docker Compose:

   docker-compose up om1 -d

2. Open the frontend:

   http://localhost:8000/
   

3. Observe frontend status:

   OM Agent Status: OFF
   

4. Check container logs:

   docker logs om1 | grep -i "websim\|event loop"

---

Actual Behavior

The container repeatedly logs:

2026-01-18 10:42:09 - WARNING - Websim tick error: This event loop is already running
/app/OM1/src/simulators/plugins/WebSim.py:589: RuntimeWarning: coroutine 'WebSim.broadcast_state' was never awaited
  loop.create_task(self.broadcast_state())
RuntimeWarning: Enable tracemalloc to get the object allocation traceback

WebSocket state updates are never delivered to the frontend.

---

Expected Behavior

• WebSim should broadcast state updates via WebSocket

• Frontend should show:

  OM Agent Status: ON
  

• No asyncio event loop errors should appear

---

Impact

• ✅ Backend core logic, APIs, and AI processing work normally
• ❌ Frontend cannot display real-time agent status
• ❌ WebSocket communication is broken

---

Root Cause

The issue originates from:

src/simulators/plugins/WebSim.py:589

loop.create_task(self.broadcast_state())

In Docker, the asyncio event loop is already running in another thread.
Calling loop.create_task() in this context raises:

This event loop is already running

As a result, the coroutine is never awaited and WebSocket broadcasting fails.

---

Suggested Fixes

Option 1: Allow nested event loops with nest_asyncio

import nest_asyncio
nest_asyncio.apply()

---

Option 2: Use thread-safe coroutine submission

asyncio.run_coroutine_threadsafe(
    self.broadcast_state(),
    loop
)

---

Option 3: Safely handle running vs non-running loops

import asyncio

def safe_create_task(coro):
    try:
        loop = asyncio.get_running_loop()
        # Loop already running
        return asyncio.ensure_future(coro)
    except RuntimeError:
        # No running loop
        loop = asyncio.new_event_loop()
        asyncio.set_event_loop(loop)
        return loop.create_task(coro)

---

Logs

Full error log

2026-01-18 10:42:01 - INFO - Initializing WebSim...
2026-01-18 10:42:01 - INFO - Starting WebSim server thread...
2026-01-18 10:42:02 - INFO - WebSim server started successfully - Open http://localhost:8000 in your browser
2026-01-18 10:42:09 - WARNING - Websim tick error: This event loop is already running
/app/OM1/src/simulators/plugins/WebSim.py:589: RuntimeWarning: coroutine 'WebSim.broadcast_state' was never awaited
  loop.create_task(self.broadcast_state())
RuntimeWarning: Enable tracemalloc to get the object allocation traceback

---

Backend Verification

API calls succeed:

docker logs om1 | grep "HTTP/1.1 200 OK"

Example output:

2026-01-18 10:41:03 - INFO - HTTP Request: POST https://api.openmind.org/api/core/openai/chat/completions "HTTP/1.1 200 OK"

AI agents are processing normally:

docker logs om1 | grep "Spot"

---

Related Files

• src/simulators/plugins/WebSim.py
• docker-compose.yml
• Dockerfile

---

Contact

bobcrypto66@gmail.com

[nan1888]

Hi! I'll work on this issue. Will use asyncio.run_coroutine_threadsafe() to fix the event loop conflict in Docker containers. PR coming soon!

[nan1888]

Update: After reviewing the codebase, I discovered this issue has already been fixed in PR #1643 (which closed duplicate issue #1621).

The fix uses asyncio.run_coroutine_threadsafe() as suggested in Option 2. The current code in WebSim.py lines 590-600 already handles the event loop conflict correctly.

This issue can be closed as a duplicate of #1621.

[Wanbogang]

@Gkkkk
already repaired.close this PR