fix: Redis集群查询+metrics修复 — 单节点连接替代ClusterClient

- crypto.go: decryptPassword 自动处理 enc: 前缀，避免调用方遗漏
- redis.go: getRedisClient 统一使用 redis.NewClient 单节点连接
  ClusterClient 拓扑发现(CLUSTER SLOTS)在Docker内网IP不可达时超时
  redisClientWrapper 全部方法改为使用 Client (不再走 ClusterClient)
  ScanAllKeys 简化为单节点 SCAN
- connection_tester.go: getRedisMetrics 统一单节点连接采集metrics
  删除 isCluster 分支的 ClusterClient 代码，避免5秒超时

backend/internal/handler/connection_tester.go

@@ -200,48 +200,29 @@ func getRedisMetrics(instance *model.Instance) map[string]interface{} {
 	}
 
 	var info string
-	if isCluster {
-		rdb := redis.NewClusterClient(&redis.ClusterOptions{
-			Addrs:       []string{addr},
-			Password:    password,
-			DialTimeout: 5 * time.Second,
-			ReadTimeout: 5 * time.Second,
-		})
-		defer rdb.Close()
+	// For metrics collection, always use single-node connection even for clusters.
+	// ClusterClient does topology discovery which can timeout if some nodes are unreachable.
+	rdb := redis.NewClient(&redis.Options{
+		Addr:         addr,
+		Password:     password,
+		DB:           0,
+		DialTimeout:  5 * time.Second,
+		ReadTimeout:  5 * time.Second,
+		WriteTimeout: 3 * time.Second,
+	})
+	defer rdb.Close()
 
-		_, err := rdb.Ping(ctx).Result()
-		if err != nil {
-			metrics["status"] = "offline"
-			return metrics
-		}
-		metrics["status"] = "online"
-		info, _ = rdb.Info(ctx, "server", "clients", "stats").Result()
-		// Latency from PING
-		start := time.Now()
-		rdb.Ping(ctx)
-		metrics["latency"] = time.Since(start).Milliseconds()
-	} else {
-		rdb := redis.NewClient(&redis.Options{
-			Addr:         addr,
-			Password:     password,
-			DB:           0,
-			DialTimeout:  5 * time.Second,
-			ReadTimeout:  5 * time.Second,
-		})
-		defer rdb.Close()
-
-		_, err := rdb.Ping(ctx).Result()
-		if err != nil {
-			metrics["status"] = "offline"
-			return metrics
-		}
-		metrics["status"] = "online"
-		info, _ = rdb.Info(ctx, "server", "clients", "stats").Result()
-		// Latency from PING
-		start := time.Now()
-		rdb.Ping(ctx)
-		metrics["latency"] = time.Since(start).Milliseconds()
+	_, err := rdb.Ping(ctx).Result()
+	if err != nil {
+		metrics["status"] = "offline"
+		return metrics
 	}
+	metrics["status"] = "online"
+	info, _ = rdb.Info(ctx, "server", "clients", "stats").Result()
+	// Latency from PING
+	start := time.Now()
+	rdb.Ping(ctx)
+	metrics["latency"] = time.Since(start).Milliseconds()
 
 	if info == "" {
 		return metrics