second phase of

• Record history hotspot information in TIDB_HOT_REGIONS #25281

There used to be a proposal pr, close it because of the change is samll and move core info to this issue, new info will be update in this issue.

• docs: add proposal to optimise tikv_region_status and tikv_region_peers #28331

Description

The existing implementation of TIKV_REGION_STATUS and TIKV_REGION_PEERS fetches all tables' regions infomation, which results in the same overhead of fetching a small table's regions infomation as a large one. To avoid this, some of the predicates can be pushed down to reduce unnecessary data transfer overhead.

1. tikv_region_status:
   The optimization idea is to add TikvRegionStatusExtractor to push down DB_NAME , TABLE_NAME, INDEX_NAME, TABLE_ID, INDEX_ID and add tikvRegionStatusRetriever to fetch regions information by region range keys, which construct start key and end key pairs from pushed down predicates.

   Task list:
   TiDB:

   • tikv_region_status extractor planner: add tikvregionstatus extractor #30649
   • tikv_region_status executor executor: add tikvregionstatus executor #30648

   PD:
   We need to support endKey in ScanRegions http interface.

   • api,pd-ctl: support endKey in ScanRegions http interface tikv/pd#4400

2. tikv_region_peers :
   The optimization idea is to add TikvRegionPeersExtractor to push down STORE_ID and REGION_ID and add tikvRegionPeersRetriever to fetch regions information from PD.

   Task list:
   TiDB:

   • add extractor for tikv_region_peers planner: add extractor for tikv_region_peers #30656
   • add retriver for tikv_region_peers executor: add retriver for tikv_region_peers  #30657

   PD: None