[tidb-x] tikv scale out stopped because pd did not generate scheduling

## Bug Report

### What did you do?
1、run sysbench with 32 threads
storage: s3://qe-testing/benchmark/sysbench_32_7000w
db : sysbench_32_7000w
tableNum : 32
tableSize : 70000000
threads : 32
cmd：sysbench --db-driver=mysql --time=0 --mysql-db=sysbench_32_7000w --mysql-host=tc-tidb.e2e-oltp-sysbench-rand-fun2-tps-8161517-1-408 --mysql-port=4000 --threads=32 --tables=32 --table-size=70000000 --mysql-ignore-errors=\"2013,1213,1205,8022,8028,9004,9007,1062\" --mysql-user=root --mysql-password= --report-interval=10 --rand-type=uniform oltp_read_only run
2、scale out tikv from 3 to 4

### What did you expect to see?
tikv scale out can succeed

### What did you see instead?
region can not balance and tikv scale out stopped

### What version of PD are you using (`pd-server -V`)?
./pd-server -V
Release Version: v9.0.0-beta.2.pre-217-g6cebb33
Edition: Enterprise
Kernel Type: Next Generation
Git Commit Hash: 6cebb3314abe4301121d39c0f0186ebb5c39d167
Git Branch: HEAD
UTC Build Time: 2025-12-11 03:58:23