Add Command Center v2: DQ/RAG integration, S3 browser, Jupyter, GPU matrix.

Mirror mo/atc-GPU layout with config/, docs/, scripts/ for Gitea deploy.

docs/runbook-vm304.md

@@ -0,0 +1,46 @@
+# Command Center VM304
+
+| Item | Value |
+|------|-------|
+| VM | MCP · Proxmox VMID **304** |
+| IP | `10.0.21.33` |
+| SSH | `root@10.0.21.33` |
+| Gitea | `http://atc-mgt01.dell-atc.lan:3001/mo/atc-agents` |
+
+## Stack
+
+- **ui** — React (Data Platform, Presentation, DQ, Knowledge Chat, S3 Storage, Jupyter link)
+- **api** — FastAPI + WebSocket + S3 browser API
+- **dq-api / rag-api** — from `mo/atc-data-quality`
+- **jupyter** — JupyterLab with S3 credentials
+- **chromadb, docling, postgres, redis, caddy**
+
+## ObjectScale S3
+
+| Item | Value |
+|------|-------|
+| Endpoint | `http://10.0.20.111:9020` |
+| Namespace | `ns1` |
+| Default user | `object_admin1` (see ECS deploy.yml) |
+| UI browse | Command Center → **Object Storage** |
+
+## Jupyter
+
+- URL: `http://10.0.21.33/jupyter/`
+- Token: `JUPYTER_TOKEN` in `.env`
+- Work dir persisted in Docker volume `jupyter_data`
+- Preconfigured: `AWS_ACCESS_KEY_ID`, `AWS_SECRET_ACCESS_KEY`, `S3_ENDPOINT`
+
+Example in notebook:
+
+```python
+import boto3, os
+s3 = boto3.client("s3", endpoint_url=os.environ["S3_ENDPOINT"])
+print(s3.list_buckets())
+```
+
+## Gitea sync
+
+```bash
+./scripts/sync-gitea.sh
+```