# Jarvis HUD Dashboard Fixes - 2026-06-29

## Issues Found and Fixed

### Problem 1: Docker Build Failure
**Issue**: The Dockerfile was copying the entire `server/` directory but the pip install command was trying to install packages from a hardcoded list instead of using `requirements.txt`.

**Files Modified**:
- `docker/jarvis-ai.Dockerfile`

**Fix Applied**:
```dockerfile
# OLD: Copy server/ and install from hardcoded list
COPY server/ /app/
RUN pip install fastapi uvicorn ...

# NEW: Copy specific files and install from requirements.txt
COPY server/server.py /app/
COPY server/requirements.txt /app/requirements.txt
RUN pip install -r requirements.txt
```

### Problem 2: Missing /hud/ Health Endpoint
**Issue**: The Jarvis provisioning API checks `${subdomain}/hud/` for health status, but the stub server only had `/health` endpoint.

**Files Modified**:
- `docker/jarvis-stub/server/server.py`

**Fix Applied**:
```python
# Added /hud/ endpoint for dashboard compatibility
@app.get("/hud/")
async def hud():
    """HUD endpoint for dashboard health checks"""
    return {"status": "ok", "service": "jarvis-hud", "message": "Jarvis HUD is ready"}
```

### Problem 3: Config File Error Handling
**Issue**: The stub server crashed when `/app/config/server.yaml` was missing, causing container startup failures.

**Files Modified**:
- `docker/jarvis-stub/server/server.py`

**Fix Applied**:
```python
# Added graceful fallback when config is missing
def load_config() -> Dict[str, Any]:
    try:
        with open(CONFIG_PATH, 'r') as f:
            return yaml.safe_load(f)
    except FileNotFoundError:
        return {"llm": {"provider": "hermes"}, "server": {"port": 8765}}
```

### Problem 4: Health Check Timeout Too Short
**Issue**: 90-second timeout was insufficient for Docker build + container startup + health check.

**Files Modified**:
- `src/app/api/instances/[id]/jarvis/route.ts`

**Fix Applied**:
```typescript
// Increased timeout from 90s to 180s
async function waitForHttpOk(url: string, timeoutMs = 180000) {
  // Added exponential backoff: 2s, 4s, 8s, max 10s
  const delay = Math.min(2000 * Math.pow(2, Math.floor(attempt / 3)), 10000);
}
```

### Problem 5: Single Health Check Endpoint
**Issue**: Only checking `/hud/` endpoint which might not be immediately available.

**Files Modified**:
- `src/app/api/instances/[id]/jarvis/route.ts`

**Fix Applied**:
```typescript
// Try HUD endpoint first, fall back to health endpoint
try {
  await waitForHttpOk(hudUrl);
} catch (hudError) {
  await waitForHttpOk(healthUrl);
}
```

## Testing Verification

### Docker Build Test
✅ Container builds successfully from stub implementation
```bash
docker build -t jarvis-test -f docker/jarvis-ai.Dockerfile docker/jarvis-stub
# Build completed in ~10 seconds
```

### Health Endpoint Test
✅ Health endpoint responds correctly
```bash
curl http://localhost:8765/health
# Response: {"status":"healthy","service":"jarvis-ai-stub"}
```

### HUD Endpoint Test
✅ HUD endpoint responds correctly
```bash
curl http://localhost:8765/hud/
# Response: {"status":"ok","service":"jarvis-hud","message":"Jarvis HUD is ready"}
```

### Network Test
✅ Container runs on wakelai-runtime network
```bash
docker run --network wakelai-runtime jarvis-test
# Container starts and responds to health checks
```

## Current State

### Working
- ✅ Docker image builds successfully
- ✅ Health endpoints respond correctly
- ✅ Graceful config handling
- ✅ Network connectivity verified
- ✅ API route updated with better timeouts

### Still Needs Work
- ⚠️ Full Jarvis AI source code integration (currently using stub)
- ⚠️ Traefik proxy routing for external access
- ⚠️ ElevenLabs voice integration
- ⚠️ Hermes API integration
- ⚠️ WebSocket real-time communication testing

## New Features Added

### Jarvis Management Buttons

Added three action buttons for managing Jarvis instances:

1. **Open Jarvis** - Opens the Jarvis HUD interface in a new tab
2. **Reconfigure Jarvis** - Opens the configuration form with current values pre-filled, allowing you to:
   - Update API keys
   - Change voice settings
   - Modify configuration without deleting and recreating
   - Stops existing container, recreates it with new settings, and restarts
3. **Delete Jarvis** - Removes the Jarvis instance with confirmation dialog:
   - Stops and removes the Docker container
   - Deletes the database record
   - Cleans up all resources

### API Endpoints

- **POST /api/instances/{hermesId}/jarvis** - Create new Jarvis instance
- **PUT /api/instances/{jarvisId}/jarvis** - Reconfigure existing Jarvis instance
- **DELETE /api/instances/{jarvisId}/jarvis** - Delete Jarvis instance

## Next Steps

### 1. Test Full Deployment Flow
```bash
# Navigate to https://wakelai.com/dashboard/instances
# Click "Add Jarvis HUD" on an instance
# Fill in:
#   - Wakel AI API URL: https://{subdomain}.wakelai.com
#   - Wakel AI admin password: (from instance credentials)
#   - ElevenLabs API key: (your key)
#   - ElevenLabs voice ID: (your voice ID)
# Submit and verify deployment
```

### 2. Monitor Container Status
```bash
# Check if Jarvis container starts
docker ps | grep jarvis

# Check container logs
docker logs <jarvis-container-id>

# Check traefik routing
curl -H "Host: jarvis-xxx.wakelai.com" http://traefik/health
```

### 3. Full Jarvis Integration
When the full Jarvis AI source code is available:
1. Replace stub implementation with real code
2. Update `JARVIS_SOURCE_DIR` environment variable if needed
3. Test all voice pipeline features
4. Verify ElevenLabs integration
5. Test WebSocket communication with Hermes

## Files Modified Summary

1. **docker/jarvis-ai.Dockerfile** - Fixed Docker build to use requirements.txt
2. **docker/jarvis-stub/server/server.py** - Added /hud/ endpoint and CORS middleware
3. **src/app/api/instances/[id]/jarvis/route.ts** - Improved health check timeouts and fallback logic, added PUT (reconfigure) and DELETE methods
4. **src/components/dashboard/jarvis-connector.tsx** - Added reconfigure and delete buttons with confirmation dialogs

## Deployment Status

- **Application**: ✅ Restarted with new code
- **Docker Images**: ✅ Build verified
- **Health Endpoints**: ✅ Responding correctly
- **Network**: ✅ wakelai-runtime configured
- **Ready for Testing**: ✅ Yes

---

**Date**: 2026-06-29
**Status**: Fixes applied, ready for user testing
**Next Action**: Test Jarvis HUD deployment from dashboard