Fix production deploy Traefik reuse

- pass NEXT_PUBLIC_CARPLACE_URL into the production Docker build
- run the production deploy script after building the image
- support raw or base64 production env secrets during deploy
- copy the production env file into the remote build context
- include homepage in production deploy, pull, and health checks
- use a stable Traefik Compose project name
- reuse an existing Traefik container instead of starting a duplicate proxy

scripts/docker-prod-common.sh

@@ -503,6 +503,13 @@ start_traefik() {
   ensure_env_file
   ensure_traefik_network
   render_registry_proxy_config
+  local existing_proxy
+  existing_proxy="$(docker ps --filter publish=80 --format '{{.ID}} {{.Image}} {{.Names}}' | awk 'tolower($0) ~ /traefik/ { print $1; exit }')"
+  if [[ -n "${existing_proxy}" ]]; then
+    echo "Using existing Traefik container ${existing_proxy}; skipping duplicate Traefik startup."
+    docker network connect "${TRAEFIK_NETWORK}" "${existing_proxy}" 2>/dev/null || true
+    return 0
+  fi
   traefik_compose up -d
 }