# Uneven Distribution Of Leader/Region: Monitoring (A)

+ The Grafana PD —> Statistics - balance page shows statistics related to load balancing


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

![right](./student-guide/diagram/25-Leader-Region-monitor-01-02.png)

------------------------------------------------------------

Uneven distribution of Leaders/Regions
- Grafana PD —> Statistics - balance page shows statistics related to load balancing
  - Store Leader Score & Store Region Score: Normally, when the Store Leader Score & Store Region Score of each TiKV node are close or the same, PD considers Leader or The region is already in equilibrium:
  - Store leader size & Store region size: Indicates the leader/region capacity of each store
  - Store Leader Count & Store Region Count: Indicates the number of Leader/Regions in each Store

Region load balancing scheduling relies mainly on two schedulers: balance-leader and balance-region. The scheduling goal for both is to spread the Region evenly across all the Stores in the cluster, but each has its own focus: balance-leader focuses on the Leader of the Region to spread the pressure of handling client request; Balance-region focuses on the peers in the Region to Distribute the pressure of storage while avoiding situations such as bursts.