Runtime Tab for Each Active Preset in Paddler Dashboard #164
Description
To complement predefined model & agent presets, it would be great if each time a preset is deployed, a new dedicated tab opens in the dashboard specifically for that active preset. This tab would provide all the information and controls a user needs to manage that preset in real time.
Key Features of the Preset Runtime Tab
Preset Overview
- Model name and version
- Number of agents and slots currently active
- Preset name and description
- Status indicator (Running, Paused, Restarting, Stopped)
Controls
- Pause / Resume preset (stop requests temporarily without shutting down agents)
- Restart preset (stop and start all agents for this preset)
- Stop / Terminate preset (completely shut down all agents and release resources)
Live Data & Monitoring
Real-time stats for this preset:
- Active agents & their slots
- Resource usage (CPU/GPU, memory) per agent
- Request throughput / latency
- Ability to scale agents up or down on the fly (optional future extension)
Logs
- Live stream of logs for all agents under this preset
- Filters for error, warning, info logs
- Search and scroll functionality
- Option to download log history for debugging or auditing
User Actions
- Quick buttons for common operations (pause, restart, stop)
- Option to open detailed agent-specific stats if needed
- Notifications or alerts when errors or high load are detected
Benefits
- Centralized Management: Users see everything about the preset in one place
- Real-time Feedback: Live data and logs help monitor performance and troubleshoot issues
- Full Control: Pause, restart, or terminate a preset without affecting other presets or agents
- Faster Iteration: Ideal for testing, scaling, and experimentation with different models
In short:
Whenever a preset is deployed, a new dashboard tab opens dedicated to that preset, showing live stats, logs, and full controls—everything a user needs to monitor, manage, and interact with that preset in real time.