Update to Domino Data Lab

{
  "description": "Deploy and monitor model API endpoints in Domino. Covers creating prediction endpoints, version management, Grafana dashboards for latency/errors/resources, alerting, and GPU inference with NVIDIA Triton. Use when deploying models as APIs, monitoring production endpoints, or debugging endpoint issues.",
  "included_files": [
    {
      "relative_path": "DEPLOY-ENDPOINT.md",
      "size_in_bytes": 12110
    },
    {
      "relative_path": "MONITORING.md",
      "size_in_bytes": 7640
    },
    {
      "relative_path": "SCALING.md",
      "size_in_bytes": 7888
    }
  ],
  "name": "domino-model-endpoints",
  "skill_md_contents": "---\nname: domino-model-endpoints\ndescription: Deploy and monitor model API endpoints in Domino. Covers creating prediction endpoints, version management, Grafana dashboards for latency/errors/resources, alerting, and GPU inference with NVIDIA Triton. Use when deploying models as APIs, monitoring production endpoints, or debugging endpoint issues.\n---\n\n# Domino Model Endpoints Skill\n\nThis skill provides comprehensive knowledge for deploying and monitoring model API endpoints in Domino Data Lab.\n\n## Key Concepts\n\n### Model Endpoints Overview\n\nDomino Model Endpoints provide:\n- REST API for model predictions\n- Automatic scaling and load balancing\n- Version management\n- Built-in monitoring with Grafana\n- Authentication via API tokens\n\n### Endpoint Lifecycle\n\n```\nTrain Model → Register → Deploy Endpoint → Monitor → Update Version\n```\n\n## Related Documentation\n\n- [domino-model-serving/SKILL.md](../domino-model-serving/SKILL.md) - REST deploy, registry v1/v2, lifecycle pitfalls\n- [API-MODEL-SERVING.md](../python-sdk/API-MODEL-SERVING.md) - modelServing routes and invoke split\n- [DEPLOY-ENDPOINT.md](./DEPLOY-ENDPOINT.md) - Creating model APIs\n- [MONITORING.md](./MONITORING.md) - Grafana, metrics, alerts\n- [SCALING.md](./SCALING.md) - GPU inference, Triton, scaling\n\n## Environment Requirements\n\n**Important:** Model APIs use the **default environment** set for your project. The environment must have the `uwsgi` Python package installed for model endpoints to work.\n\n### Required Package\n```dockerfile\n# Add to your environment's Dockerfile instructions\nRUN pip install uwsgi\n```\n\nOr in requirements.txt:\n```\nuwsgi\n```\n\n### Setting Default Environment\n1. Go to **Project Settings** → **Execution Preferences**\n2. Set the **Default Environment** that includes `uwsgi`\n3. This environment will be used for all Model API deployments\n\n## Quick Start\n\n### 1. Create Endpoint Function\n\n```python\n# model.py\ndef predict(features):\n    \"\"\"\n    Domino calls this function for predictions.\n\n    Args:\n        features: Input data (dict, list, or primitive)\n\n    Returns:\n        JSON-serializable prediction result\n    \"\"\"\n    import pickle\n\n    # Load model (cached after first call)\n    with open('model.pkl', 'rb') as f:\n        model = pickle.load(f)\n\n    prediction = model.predict([features])\n    return {\"prediction\": prediction.tolist()}\n```\n\n### 2. Deploy via Domino UI\n\n1. Go to **Publish** → **Model APIs**\n2. Click **New Model**\n3. Configure:\n   - Name: `my-classifier`\n   - File: `model.py`\n   - Function: `predict`\n   - Environment: Select compute environment\n4. Click **Publish**\n\n### 3. Call the Endpoint\n\n```bash\ncurl -X POST \\\n  -H \"Content-Type: application/json\" \\\n  -H \"Authorization: Bearer YOUR_API_TOKEN\" \\\n  -d '{\"data\": {\"features\": [1.0, 2.0, 3.0]}}' \\\n  https://your-domino.com/models/abc123/latest/model\n```\n\n## Environment Variables\n\nWhen calling endpoints from apps:\n\n| Variable | Description |\n|----------|-------------|\n| `MODEL_API_URL` | Full endpoint URL |\n| `MODEL_API_TOKEN` | Bearer token for authentication |\n\n## Key Metrics to Monitor\n\n| Metric | Target |\n|--------|--------|\n| Latency P50 | < 100ms |\n| Latency P99 | < 500ms |\n| Error Rate | < 1% |\n| CPU Usage | < 80% |\n| Memory | Stable (no growth) |\n\n## Documentation Links\n\nOpenAPI and route discovery: [API-SPECS.md](../domino-api-intro/API-SPECS.md) (public routes section).\n"
}