[improve](routine load) ensure load balance after scaling up BE nodes (…

…#44693)

We scaling up from three BE nodes to five BE nodes, but from the
monitoring perspective, only the previous three nodes have written
traffic.

![image](https://github.com/user-attachments/assets/947fe8a9-745d-4c37-93d4-15b0e27f12fc)

This pr aims to ensure load balance after scaling up BE nodes.

fe/fe-core/src/main/java/org/apache/doris/load/routineload/RoutineLoadManager.java

@@ -487,6 +487,7 @@ public long getAvailableBeForTask(long jobId, long previousBeId) throws LoadExce
         // check if be has idle slot
         readLock();
         try {
+            updateBeIdToMaxConcurrentTasks();
             Map<Long, Integer> beIdToConcurrentTasks = getBeCurrentTasksNumMap();
             int previousBeIdleTaskNum = 0;