[FEAT] 調整生產環境部署流程的容器健康檢查等待邏輯與清理策略

1. 修改 .gitea/workflows/deploy-prod.yaml 的 Step 1，將 docker-compose --wait 改為手動 shell 迴圈，動態輪詢 star-cloud-mysql、star-cloud-redis、star-cloud-emqx 的健康狀態，並在超時 (120秒) 時輸出日誌以提升部署強固性。
2. 調整 Step 5 - Cleanup，將 docker image prune -f 限制為只清理超過一週 (168h) 的過期映像檔，以保留 Docker 建置快取。

.gitea/workflows/deploy-prod.yaml

@@ -33,7 +33,28 @@ jobs:
 
       - name: Step 1 - Build & Deploy Containers
         run: |
-          WWWGROUP=1001 WWWUSER=1001 docker-compose up -d --build --wait
+          WWWGROUP=1001 WWWUSER=1001 docker-compose up -d --build
+
+          for container in star-cloud-mysql star-cloud-redis star-cloud-emqx; do
+            echo ">>> Waiting for ${container} to become healthy..."
+
+            for i in $(seq 1 60); do
+              status=$(docker inspect -f '{{if .State.Health}}{{.State.Health.Status}}{{else}}{{.State.Status}}{{end}}' "${container}" 2>/dev/null || echo "missing")
+
+              if [ "${status}" = "healthy" ]; then
+                echo ">>> ${container} is healthy."
+                break
+              fi
+
+              if [ "${i}" -eq 60 ]; then
+                echo ">>> ${container} did not become healthy. Current status: ${status}"
+                docker logs --tail 80 "${container}" || true
+                exit 1
+              fi
+
+              sleep 2
+            done
+          done
 
       - name: Step 2 - Deploy Code
         run: |
@@ -75,4 +96,4 @@ jobs:
           echo "✅ MQTT Worker 已重啟"
 
       - name: Step 5 - Cleanup
-        run: docker image prune -f
+        run: docker image prune -f --filter "until=168h"