Real-time GPU and system resource monitoring for ComfyUI with per-node performance tracking.
-
📊 Real-time Resource Monitoring
- GPU utilization, VRAM usage, temperature, and power draw
- CPU usage and RAM statistics
- Updates every second via WebSocket
-
🎯 Per-Node Performance Tracking (Unique Feature!)
- Track CPU and GPU usage for each individual node
- Measure execution time per node
- Identify workflow bottlenecks
- See which nodes consume the most resources
-
🎨 Clean, Integrated UI
- Embedded monitoring display in the node
- Expandable node performance list
- Real-time updates without workflow execution
- Professional dark theme design
-
🚀 Minimal Performance Impact
- Thread-safe background monitoring (<0.5% CPU)
- No impact on ComfyUI execution
- Smart resource sampling
cd ComfyUI/custom_nodes/
git clone https://github.com/ComfyAssets/ComfyUI-KikoStats.git
cd ComfyUI-KikoStats
pip install -r requirements.txt- Download/copy the
ComfyUI-KikoStatsfolder to yourComfyUI/custom_nodes/directory - Install dependencies:
pip install nvidia-ml-py psutil
Coming soon - will be available through ComfyUI Manager
- Restart ComfyUI after installation
- In ComfyUI, add a new node and search for "Resource Monitor"
- Find it under the ComfyAssets category
- Click "Start Monitoring" button to see real-time stats
- Run any workflow to see per-node performance data
Real-time Stats:
🖥️ GPU: 45%
VRAM: 1024MB / 24576MB
Temp: 37°C
💻 CPU: 12.5%
RAM: 3831MB / 64197MB
Per-Node Performance (after workflow execution):
📊 Node Performance
KSampler: 12.8% CPU, 87.3% GPU
(15.7s)
VAE Decode: 15.2% CPU, 45.8% GPU
(2.3s)
Load Checkpoint: 8.2% CPU, 12.1% GPU
(1.2s)
- Restart ComfyUI completely (most common fix)
- Check ComfyUI console for error messages during startup
- Verify the folder is in the correct location:
ComfyUI/custom_nodes/ComfyUI-KikoStats/ - Check that dependencies are installed:
pip install nvidia-ml-py psutil
- Requires NVIDIA GPU with drivers installed
- Install nvidia-ml-py:
pip install nvidia-ml-py - Check that NVIDIA drivers are working:
nvidia-smi
- Install psutil:
pip install psutil
Category: ComfyAssets
Inputs:
update_interval(FLOAT): Monitoring refresh rate in seconds (0.1-60.0)display_mode(CHOICE): Output format - "text", "json", or "both"enable_gpu(BOOLEAN): Enable GPU monitoring (optional)enable_system(BOOLEAN): Enable system monitoring (optional)
Outputs:
stats(STRING): Formatted monitoring statisticsjson_data(STRING): JSON data for chaining to other nodes
- Core: Python 3.7+
- GPU Monitoring: nvidia-ml-py (for NVIDIA GPUs)
- System Monitoring: psutil
- ComfyUI: Compatible with recent versions
- CPU Overhead: <0.5% typical usage
- Memory: Minimal impact with circular buffers
- Thread Safety: Background monitoring doesn't block ComfyUI execution
| Feature | KikoStats | Crystools | Elegant Monitor |
|---|---|---|---|
| Real-time monitoring | ✅ | ✅ | ✅ |
| Per-node tracking | ✅ | ❌ | ❌ |
| GPU stats | ✅ | ✅ | ✅ |
| CPU/RAM stats | ✅ | ✅ | ✅ |
| Embedded UI | ✅ | ❌ | ✅ |
| WebSocket updates | ✅ | ✅ | ❌ |
| Execution timing | ✅ | ❌ | ❌ |
MIT License - see LICENSE file for details.