[Dynamic Regions] PD does not schedule region-balance when there's moderate read-write traffic #14214
Description
Bug Report
What version of TiKV are you using?
release-6.6
What operating system and CPU are you using?
Steps to reproduce
- Sysbench prepare 150GB data per TiKV
- scale-out a TiKV on a 3 TiKV cluster , wait for 5 minutes
- run Sysbench read-write workload
What did you expect?
The region balance speed is similar with or without sysbench workload
What did happened?
The region balance operator is not scheduled when there's moderate read-write workload.