# Hotspot Region Are Unevenly Distributed (C)


![left](./student-guide/diagram/25-Region-solution-01-01.png)

------------------------------------------------------------
solutions

- Slow scheduling of hotspots
From the metrics of PD, it can be seen that there are quite a few hot regions, but the scheduling speed cannot keep up, and the hot regions cannot be dispersed in time.
  - Increase the hot-region-schedule-limit and reduce the limit quotas of other schedulers to speed up hotspot scheduling
  - A small reduction in the hot-region-cache-hits-threshold allows the PD to react more quickly to changes in traffic.

- A single region forms a hot spot
Generally, it means that a large number of request frequently scan a small table, which can be seen from the business perspective or the hot information of the TiDB Dashboard traffic visualization statistics.
  - You can reduce load base split related parameters such as split.qps-threshold and split.byte-threshold
  - Manually add a split-region schedule to split such a region
    - First use TiDB Dashboard traffic visualization to confirm the hotspot region
    - Use the pd-ctl operator add split-region {region_id} —policy=to manually split the target region