Added traefik for production inference through tunnel

compose/prod/docker-compose.inference.yml

@@ -11,6 +11,7 @@ services:
             - driver: nvidia
               count: all
               capabilities: [gpu]
+    restart: unless-stopped
     environment:
       - INFERENCE_HTTP_HOST=0.0.0.0
       - INFERENCE_HTTP_PORT=8001
@@ -19,9 +20,19 @@ services:
       - PYTHONPATH=/app
       - HF_HOME=/root/.cache/huggingface
       - HF_HUB_OFFLINE=1
-    ports:
-      - '0.0.0.0:58001:8001'
     volumes:
       - ../../:/app
       - ../../models:/app/models
       - ../../hf_cache:/root/.cache/huggingface
+    labels:
+      - "traefik.enable=true"
+      - "traefik.http.routers.llm.rule=Host(`${LLM_DOMAIN_NAME}`)"
+      - "traefik.http.routers.llm.entrypoints=web"
+      - "traefik.http.services.llm.loadbalancer.server.port=${LLM_PORT}"
+      - "traefik.docker.network=proxy-net"
+    networks:
+      - proxy-net
+
+networks:
+  proxy-net:
+    external: true