Back to Tidb

TiDB HTTP API

docs/tidb_http_api.md

26.3.732.9 KB
Original Source

TiDB HTTP API

TiDBIP is the ip of the TiDB server. 10080 is the default status port, and you can edit it in tidb.toml when starting the TiDB server.

  1. Get the current status of TiDB, including the connections, version, and git_hash.

    shell
    curl http://{TiDBIP}:10080/status
    
    shell
    $curl http://127.0.0.1:10080/status
    {
        "connections": 0,
        "git_hash": "f572e33854e1c0f942f031e9656d0004f99995c6",
        "version": "5.7.25-TiDB-v2.1.0-rc.3-355-gf572e3385-dirty",
        "status":{
          "init_stats_percentage":100
        }
    }
    
  2. Get all metrics of TiDB

    shell
    curl http://{TiDBIP}:10080/metrics
    
  3. Get the metadata of all regions

    shell
    curl http://{TiDBIP}:10080/regions/meta
    
    shell
    $curl http://127.0.0.1:10080/regions/meta
    [
        {
            "leader": {
                "id": 5,
                "store_id": 1
            },
            "peers": [
                {
                    "id": 5,
                    "store_id": 1
                }
            ],
            "region_epoch": {
                "conf_ver": 1,
                "version": 2
            },
            "region_id": 4
        }
    ]
    
  4. Get the table/index of hot regions

    shell
    curl http://{TiDBIP}:10080/regions/hot
    
    shell
    $curl http://127.0.0.1:10080/regions/hot
    {
      "read": [
    
      ],
      "write": [
        {
          "db_name": "sbtest1",
          "table_name": "sbtest13",
          "index_name": "",
          "flow_bytes": 220718,
          "max_hot_degree": 12,
          "region_count": 1
        }
      ]
    }
    
  5. Get the information of a specific region by ID

    shell
    curl http://{TiDBIP}:10080/regions/{regionID}
    
    shell
    $curl http://127.0.0.1:10080/regions/4001
    {
        "end_key": "dIAAAAAAAAEk",
        "frames": [
            {
                "db_name": "test",
                "is_record": true,
                "table_id": 286,
                "table_name": "t1"
            }
        ],
        "region_id": 4001,
        "start_key": "dIAAAAAAAAEe"
    }
    
  6. Get regions Information from db.table

    shell
    curl http://{TiDBIP}:10080/tables/{db}/{table}/regions
    
    shell
    $curl http://127.0.0.1:10080/tables/test/t1/regions
    {
        "id": 286,
        "indices": [],
        "name": "t1",
        "record_regions": [
            {
                "leader": {
                    "id": 4002,
                    "store_id": 1
                },
                "peers": [
                    {
                        "id": 4002,
                        "store_id": 1
                    }
                ],
                "region_epoch": {
                    "conf_ver": 1,
                    "version": 83
                },
                "region_id": 4001
            }
        ]
    }
    
  7. Get schema Information about all db

    shell
    curl http://{TiDBIP}:10080/schema
    
    shell
    $curl http://127.0.0.1:10080/schema
    [
        {
            "charset": "utf8mb4",
            "collate": "utf8mb4_bin",
            "db_name": {
                "L": "test",
                "O": "test"
            },
            "id": 266,
            "state": 5
        },
        .
        .
        .
    ]
    
  8. Get schema Information about db

    shell
    curl http://{TiDBIP}:10080/schema/{db}
    
    shell
    curl http://{TiDBIP}:10080/schema/{db}?id_name_only=true
    [
    {
     "id": 119,
     "name": {
      "O": "t1",
      "L": "t1"
     }
    },
    {
     "id": 125,
     "name": {
      "O": "t2",
      "L": "t2"
     }
    }
    ]
    
  9. Get schema Information about db.table, and you can get schema info by tableID (tableID is the unique identifier of table in TiDB)

    shell
    curl http://{TiDBIP}:10080/schema/{db}/{table}
    
    curl http://{TiDBIP}:10080/schema?table_id={tableID}
    
    curl http://{TiDBIP}:10080/schema?table_ids={tableID,...}
    
  10. Get database information, table information and tidb info schema version by tableID.

    shell
    curl http://{TiDBIP}:10080/db-table/{tableID}
    
  11. Get MVCC Information of the key with a specified handle ID

    shell
    curl http://{TiDBIP}:10080/mvcc/key/{db}/{table}/{handle}
    
    shell
    $curl http://127.0.0.1:10080/mvcc/key/test/t/1
    {
        "key": "74800000000000006E5F728000000000000001",
        "region_id": 4,
        "value": {
            "info": {
                "writes": [
                    {
                        "start_ts": 448662063415296001,
                        "commit_ts": 448662063415296003,
                        "short_value": "gAABAAAAAQEAAQ=="
                    }
                ]
            }
        }
    }
    

    If the handle is clustered, specify the primary key column values in the query string

    shell
    $curl "http://{TiDBIP}:10080/mvcc/key/{db}/{table}?${c1}={v1}&${c2}=${v2}"
    
    shell
    $curl "http://127.0.0.1:10080/mvcc/key/test/t?a=aaa&b=2020-01-01"
    {
        "key": "7480000000000000365F72016161610000000000FA0419A5420000000000",
        "region_id": 52,
        "value": {
            "info": {
                "writes": [
                    {
                        "type": 1,
                        "start_ts": 423158426542538752,
                        "commit_ts": 423158426543587328
                    },
                    {
                        "start_ts": 423158426542538752,
                        "commit_ts": 423158426543587328,
                        "short_value": "gAACAAAAAQMDAAQAYWFhZA=="
                    }
                ],
                "values": [
                    {
                        "start_ts": 423158426542538752,
                        "value": "gAACAAAAAQMDAAQAYWFhZA=="
                    }
                ]
            }
        }
    }
    

    Hint: The meaning of the MVCC operation type:

    protobuf
    enum Op {
        Put = 0;
        Del = 1;
        Lock = 2;
        Rollback = 3;
        // insert operation has a constraint that key should not exist before.
        Insert = 4;
        PessimisticLock = 5;
        CheckNotExists = 6;
    }
    

    Hint: On a partitioned table, use the table(partition) pattern as the table name, t1(p1) for example:

    shell
    $curl http://127.0.0.1:10080/mvcc/key/test/t1(p1)/1
    

    Hint: The method to convert the Hex format key returned by TiDB API into the format recognized by tikv-ctl.

    Step 1: Get the hex format of the key you need. For example you could find the key by the following TiDB API.

    shell
       $curl http://127.0.0.1:10080/mvcc/key/test/t1/1
       {
       "key": "7480000000000008C65F728000000000000001",
       "region_id": 10,
       "value": {
           "info": {
               "writes": [
                   {
                       "start_ts": 445971968923271174,
                       "commit_ts": 445971968923271175,
                       "short_value": "gAACAAAAAgMIAAkAc2hpcmx5YTQi"
                   },
                   {
                       "start_ts": 445971959499980803,
                       "commit_ts": 445971959499980804,
                       "short_value": "gAACAAAAAgMIAAkAc2hpcmx5YTQL"
                   }
               ]
           }
       }
       }
    

    Step 2: Convert the key from hex format to escaped format with tikv-ctl

    shell
     ./tikv-ctl --to-escaped '7480000000000008C65F728000000000000001'
     t\200\000\000\000\000\000\010\306_r\200\000\000\000\000\000\000\001
    

    Step 3: Encode the key to make it memcomparable in tikv with tikv-ctl

    shell
    ./tikv-ctl --encode 't\200\000\000\000\000\000\010\306_r\200\000\000\000\000\000\000\001'
    7480000000000008FFC65F728000000000FF0000010000000000FA
    

    Step 4: Convert the key from hex format to escaped format again since most tikv-ctl commands only accept keys in escaped format while the --encode command outputs the key in hex format.

    shell
    ./tikv-ctl --to-escaped '7480000000000008FFC65F728000000000FF0000010000000000FA'
    t\200\000\000\000\000\000\010\377\306_r\200\000\000\000\000\377\000\000\001\000\000\000\000\000\372
    

    Step 5: Add a prefix "z" to the key. Then the key can be recognized by tikv-ctl. For example, use the following command to scan from tikv.

    shell
    ./tikv-ctl  --host "<tikv_ip>:<port>" scan --from 'zt\200\000\000\000\000\000\010\377\306_r\200\000\000\000\000\377\000\000\001\000\000\000\000\000\372' --limit 5 --show-cf write,lock,default 
       key: zt\200\000\000\000\000\000\010\377\306_r\200\000\000\000\000\377\000\000\001\000\000\000\000\000\372
        write cf value: start_ts: 445971968923271174 commit_ts: 445971968923271175 short_value: 800002000000020308000900736869726C79613422
        write cf value: start_ts: 445971959499980803 commit_ts: 445971959499980804 short_value: 800002000000020308000900736869726C7961340B
    
       key: zt\200\000\000\000\000\000\010\377\306_r\200\000\000\000\000\377\000\000\002\000\000\000\000\000\372
        write cf value: start_ts: 445971960836390913 commit_ts: 445971960836390914 short_value: 80000200000002030500060073686972340B
    
       key: zt\200\000\377\377\377\377\377\377\373_r\200\000\000\000\000\377\000\000\003\000\000\000\000\000\372
        write cf value: r_type: Del start_ts: 444068474890485761 commit_ts: 444068474890485762
    
       key: zt\200\000\377\377\377\377\377\377\373_r\200\000\000\000\000\377\000\000\005\000\000\000\000\000\372
        write cf value: r_type: Del start_ts: 444068474929545217 commit_ts: 444068474929545218
    
       key: zt\200\000\377\377\377\377\377\377\373_r\200\000\000\000\000\377\000\000\007\000\000\000\000\000\372
        write cf value: r_type: Del start_ts: 444068474981974017 commit_ts: 444068474981974018
    
  12. Get MVCC Information of the first key in the table with a specified start ts

    shell
    curl http://{TiDBIP}:10080/mvcc/txn/{startTS}/{db}/{table}
    
    shell
    $curl http://127.0.0.1:10080/mvcc/txn/405179368526053377/test/t1
    {
        "info": {
            "writes": [
                {
                    "commit_ts": 405179368526053380,
                    "short_value": "CAICAkE=",
                    "start_ts": 405179368526053377
                }
            ]
        },
        "key": "dIAAAAAAAAEzX3KAAAAAAAAAAQ=="
    }
    
  13. Get MVCC Information by a hex value

    shell
    curl http://{TiDBIP}:10080/mvcc/hex/{hexKey}
    
  14. Get MVCC Information of a specified index key, argument example: column_name_1=column_value_1&column_name_2=column_value2...

    shell
    curl "http://{TiDBIP}:10080/mvcc/index/{db}/{table}/{index}/{handle}?${c1}={v1}&${c2}=${v2}"
    

    Hint: For the index column which column type is timezone dependent, e.g. timestamp, convert its value to UTC timezone.

    shell
    $curl "http://127.0.0.1:10080/mvcc/index/test/t1/idx/1?a=A"
    {
        "info": {
            "writes": [
                {
                    "commit_ts": 405179523374252037,
                    "short_value": "MA==",
                    "start_ts": 405179523374252036
                }
            ]
        }
    }
    

    Hint: On a partitioned table, use the table(partition) pattern as the table name, t1(p1) for example:

    shell
    $curl "http://127.0.0.1:10080/mvcc/index/test/t1(p1)/idx/1?a=A"
    

    If the handle is clustered, also specify the primary key column values in the query string

    shell
    $curl "http://{TiDBIP}:10080/mvcc/index/{db}/{table}/{index}?${c1}={v1}&${c2}=${v2}"
    
    shell
    $curl "http://127.0.0.1:10080/mvcc/index/test/t/idx?a=1.1&b=111&c=1"
    {
        "key": "74800000000000003B5F69800000000000000203800000000000000105BFF199999999999A013131310000000000FA",
        "region_id": 59,
        "value": {
            "info": {
                "writes": [
                    {
                        "start_ts": 424752858505150464,
                        "commit_ts": 424752858506461184,
                        "short_value": "AH0B"
                    }
                ],
                "values": [
                    {
                         "start_ts": 424752858505150464,
                         "value": "AH0B"
                    }
                ]
            }
        }
    }
    
    
  15. Scatter regions of the specified table, add a scatter-range scheduler for the PD and the range is same as the table range.

    shell
    curl http://{TiDBIP}:10080/tables/{db}/{table}/scatter
    

    Hint: On a partitioned table, use the table(partition) pattern as the table name, test(p1) for example.

    Note: The scatter-range scheduler may conflict with the global scheduler, do not use it for long periods on the larger table.

  16. Stop scatter the regions, disable the scatter-range scheduler for the specified table.

    shell
    curl http://{TiDBIP}:10080/tables/{db}/{table}/stop-scatter
    

    Hint: On a partitioned table, use the table(partition) pattern as the table name, test(p1) for example.

  17. Get TiDB server settings

    shell
    curl http://{TiDBIP}:10080/settings
    
  18. Get TiDB server information.

    shell
    curl http://{TiDBIP}:10080/info
    
    shell
    $curl http://127.0.0.1:10080/info
    {
        "ddl_id": "f7e73ed5-63b4-4cb4-ba7c-42b32dc74e77",
        "git_hash": "f572e33854e1c0f942f031e9656d0004f99995c6",
        "ip": "",
        "is_owner": true,
        "lease": "45s",
        "listening_port": 4000,
        "status_port": 10080,
        "version": "5.7.25-TiDB-v2.1.0-rc.3-355-gf572e3385-dirty"
    }
    
  19. Get TiDB cluster all servers information.

    shell
    curl http://{TiDBIP}:10080/info/all
    
    shell
    $curl http://127.0.0.1:10080/info/all
    {
        "servers_num": 2,
        "owner_id": "29a65ec0-d931-4f9e-a212-338eaeffab96",
        "is_all_server_version_consistent": true,
        "all_servers_info": {
            "29a65ec0-d931-4f9e-a212-338eaeffab96": {
                "version": "5.7.25-TiDB-v4.0.0-alpha-669-g8f2a09a52-dirty",
                "git_hash": "8f2a09a52fdcaf9d9bfd775d2c6023f363dc121e",
                "ddl_id": "29a65ec0-d931-4f9e-a212-338eaeffab96",
                "ip": "",
                "listening_port": 4000,
                "status_port": 10080,
                "lease": "45s",
                "binlog_status": "Off"
            },
            "cd13c9eb-c3ee-4887-af9b-e64f3162d92c": {
                "version": "5.7.25-TiDB-v4.0.0-alpha-669-g8f2a09a52-dirty",
                "git_hash": "8f2a09a52fdcaf9d9bfd775d2c6023f363dc121e",
                "ddl_id": "cd13c9eb-c3ee-4887-af9b-e64f3162d92c",
                "ip": "",
                "listening_port": 4001,
                "status_port": 10081,
                "lease": "45s",
                "binlog_status": "Off"
            }
        }
    }
    
  20. Enable/Disable TiDB server general log

    shell
    curl -X POST -d "tidb_general_log=1" http://{TiDBIP}:10080/settings
    curl -X POST -d "tidb_general_log=0" http://{TiDBIP}:10080/settings
    
  21. Change TiDB server log level

    shell
    curl -X POST -d "log_level=debug" http://{TiDBIP}:10080/settings
    curl -X POST -d "log_level=info" http://{TiDBIP}:10080/settings
    
  22. Change TiDB DDL slow log threshold

    The unit is millisecond.

    shell
    curl -X POST -d "ddl_slow_threshold=300" http://{TiDBIP}:10080/settings
    
  23. Get the column value by an encoded row and some information that can be obtained from a column of the table schema information.

    Argument example: rowBin=base64_encoded_row_value

    shell
    curl http://{TiDBIP}:10080/tables/{colID}/{colFlag}/{colLen}?rowBin={val}
    

    Hint: For the column which field type is timezone dependent, e.g. timestamp, convert its value to UTC timezone.

  24. Resign the ddl owner, let tidb start a new ddl owner election.

    shell
    curl -X POST http://{TiDBIP}:10080/ddl/owner/resign
    

    Note: If you request a TiDB that is not ddl owner, the response will be This node is not a ddl owner, can't be resigned.

  25. Get the TiDB DDL job history information.

    shell
    curl http://{TiDBIP}:10080/ddl/history
    

    Note: When the DDL history is very very long, system table may containing too many jobs. This interface will get a maximum of 2048 history ddl jobs by default. If you want get more jobs, consider adding start_job_id and limit.

  26. Get count {number} TiDB DDL job history information.

    shell
    curl http://{TiDBIP}:10080/ddl/history?limit={number}
    
  27. Get count {number} TiDB DDL job history information, start with job {id}

    shell
    curl "http://{TIDBIP}:10080/ddl/history?start_job_id={id}&limit={number}"
    
  28. Download TiDB debug info

    shell
    curl http://{TiDBIP}:10080/debug/zip?seconds=60 --output debug.zip
    

    zip file will include:

    • Go heap pprof(after GC)
    • Go cpu pprof(10s)
    • Go mutex pprof
    • Full goroutine
    • TiDB config and version

    Param:

    • seconds: profile time(s), default is 10s.
  29. Get statistics data of specified table.

    shell
    curl http://{TiDBIP}:10080/stats/dump/{db}/{table}
    
  30. Get statistics data of specific table and timestamp.

    shell
    curl http://{TiDBIP}:10080/stats/dump/{db}/{table}/{yyyyMMddHHmmss}
    
    shell
    curl http://{TiDBIP}:10080/stats/dump/{db}/{table}/{yyyy-MM-dd HH:mm:ss}
    
  31. Resume the binlog writing when Pump is recovered.

    shell
    curl http://{TiDBIP}:10080/binlog/recover
    

    Return value:

    • timeout, return status code: 400, message: timeout

    • If it returns normally, status code: 200, message example:

      text
      {
        "Skipped": false,
        "SkippedCommitterCounter": 0
      }
      

      Skipped: false indicates that the current binlog is not in the skipped state, otherwise, it is in the skipped state SkippedCommitterCounter: Represents how many transactions are currently being committed in the skipped state. By default, the API will return after waiting until all skipped-binlog transactions are committed. If this value is greater than 0, it means that you need to wait until them are committed .

    Param:

    • op=nowait: return after binlog status is recoverd, do not wait until the skipped-binlog transactions are committed.
    • op=reset: reset SkippedCommitterCounter to 0 to avoid the problem that SkippedCommitterCounter is not cleared due to some unusual cases.
    • op=status: Get the current status of binlog recovery.
  32. Enable/disable async commit feature

    shell
    curl -X POST -d "tidb_enable_async_commit=1" http://{TiDBIP}:10080/settings
    curl -X POST -d "tidb_enable_async_commit=0" http://{TiDBIP}:10080/settings
    
  33. Enable/disable one-phase commit feature

    shell
    curl -X POST -d "tidb_enable_1pc=1" http://{TiDBIP}:10080/settings
    curl -X POST -d "tidb_enable_1pc=0" http://{TiDBIP}:10080/settings
    
  34. Enable/disable the mutation checker

    shell
    curl -X POST -d "tidb_enable_mutation_checker=1" http://{TiDBIP}:10080/settings
    curl -X POST -d "tidb_enable_mutation_checker=0" http://{TiDBIP}:10080/settings
    
  35. Get/Set the size of the Ballast Object

    shell
    # get current size of the ballast object
    curl -v http://{TiDBIP}:10080/debug/ballast-object-sz
    # reset the size of the ballast object (2GB in this example)
    curl -v -X POST -d "2147483648" http://{TiDBIP}:10080/debug/ballast-object-sz
    
  36. Set deadlock history table capacity

    shell
    curl -X POST -d "deadlock_history_capacity={number}" http://{TiDBIP}:10080/settings
    
  37. Set whether deadlock history (DEADLOCKS) collect retryable deadlocks

    shell
    curl -X POST -d "deadlock_history_collect_retryable={bool_val}" http://{TiDBIP}:10080/settings
    
  38. Set transaction_id to digest mapping minimum duration threshold, only transactions which last longer than this threshold will be collected into TRX_SUMMARY.

    shell
    curl -X POST -d "transaction_id_digest_min_duration={number}" http://{TiDBIP}:10080/settings
    

    Unit of duration here is ms.

  39. Set transaction summary table (TRX_SUMMARY) capacity

    shell
    curl -X POST -d "transaction_summary_capacity={number}" http://{TiDBIP}:10080/settings
    
  40. The commands are used to handle smooth upgrade mode(refer to the TiDB Smooth Upgrade for details) operations. We can send these upgrade operations to the cluster. The operations here include start, finish and show.

    shell
    curl -X POST http://{TiDBIP}:10080/upgrade/{op}
    
    shell
    $curl -X POST http://{TiDBIP}:10080/upgrade/start
    "success!"
    
  41. Set split & scatter regions concurrency before ingest, and ingest request concurrency. Value ranges:

    • max-batch-split-ranges: [1, 9223372036854775807], default 2048
    • max-split-ranges-per-sec: [0, 9223372036854775807], default 0 (no limit)
    • max-ingest-per-sec: [0, 9223372036854775807], default 0 (no limit)
    • max-ingest-inflight: [0, 9223372036854775807], default 0 (no limit)
    shell
    curl http://{TiDBIP}:10080/ingest/max-batch-split-ranges
    curl http://{TiDBIP}:10080/ingest/max-split-ranges-per-sec
    curl http://{TiDBIP}:10080/ingest/max-ingest-per-sec
    curl http://{TiDBIP}:10080/ingest/max-ingest-inflight
    
    shell
    curl http://{TiDBIP}:10080/ingest/max-batch-split-ranges -X POST -d "{\"value\": 1024}"
    curl http://{TiDBIP}:10080/ingest/max-split-ranges-per-sec -X POST -d "{\"value\": 16}"
    curl http://{TiDBIP}:10080/ingest/max-ingest-per-sec -X POST -d "{\"value\": 0.5}"
    curl http://{TiDBIP}:10080/ingest/max-ingest-inflight -X POST -d "{\"value\": 2}"
    
  42. Get TiDB transaction GC states:

    shell
    curl http://{TiDBIP}:10080/txn-gc-states
    

Test-only APIs (enableTestAPI failpoint)

These APIs are only registered when the enableTestAPI failpoint is enabled.

Enable

Build tidb-server with the failpoint enabled, then run with:

shell
GO_FAILPOINTS="github.com/pingcap/tidb/pkg/server/enableTestAPI=return" ./bin/tidb-server

Delete row key or index key

You can use those APIs to mock dangling row or index keys case for test.

Row key:

shell
curl -X POST "http://{TiDBIP}:10080/test/delete/rowkey/{db}/{table}?handle={intHandle}"
# For clustered common handle tables:
curl -X POST "http://{TiDBIP}:10080/test/delete/rowkey/{db}/{table}?{pkCol}={pkVal}[&{pkCol2}={pkVal2}...]"

Index key:

shell
curl -X POST "http://{TiDBIP}:10080/test/delete/indexkey/{db}/{table}/{index}?handle={intHandle}&{idxCol}={idxVal}[&{idxCol2}={idxVal2}...]"
# For clustered common handle tables:
curl -X POST "http://{TiDBIP}:10080/test/delete/indexkey/{db}/{table}/{index}?{idxCol}={idxVal}[&{idxCol2}={idxVal2}...]"

APIs unique to TiDB-X

Run ADMIN CHECK for an index

shell
# curl -XPOST "http://{TiDBIP}:10080/ddl/check/{db}/{table}/{index}"
{
 "check_sql": "admin check index `test`.`t` `primary`",
 "db": "test",
 "index": "primary",
 "result": "success",
 "table": "t"
}

APIs unique to TiDB-X SYSTEM keyspace

These APIs are registered only on TiDB instances running in the TiDB-X SYSTEM keyspace. All examples in this section assume {TiDBIP}:10080 belongs to a TiDB process in that SYSTEM keyspace.

The /dxf/... APIs are for DXF (Distributed eXecution Framework) operator observability and emergency runtime tuning. In TiDB-X, DXF runs as a shared SYSTEM-keyspace service for resource-intensive work such as IMPORT INTO and distributed add index, so some APIs are called on a SYSTEM-keyspace TiDB process but target a user keyspace parameter.

Get the DXF schedule status

This API returns the scheduler view used by cluster controllers to decide how many DXF worker nodes are needed. It includes the number of scheduled tasks, current and required TiDB/TiKV worker counts, busy DXF nodes, and temporary scheduler flags such as pause_scale_in.

Usage:

shell
curl http://{TiDBIP}:10080/dxf/schedule/status

Parameters: none.

Example response:

json
{
 "version": 1,
 "task_queue": {
  "scheduled_count": 2
 },
 "tidb_worker": {
  "cpu_count": 10,
  "required_count": 2,
  "current_count": 3,
  "busy_nodes": [
   {
    "id": "192.168.1.10:5000",
    "is_owner": true
   }
  ]
 },
 "tikv_worker": {
  "required_count": 2
 },
 "flags": {
  "pause_scale_in": {
   "enabled": true,
   "ttl": 3600000000000,
   "expire_time": "2025-08-22T16:58:23.092864+08:00"
  }
 }
}

Pause or resume DXF worker scale-in

This API writes a temporary DXF schedule flag. Use it when scale-in conflicts with DXF subtask scheduling, for example when subtasks keep being moved away from nodes selected for scale-in and task completion becomes slower. pause_scale_in asks the cluster controller to stop scaling in DXF workers for a limited time. resume_scale_in clears that request.

Usage:

shell
curl -X POST "http://{TiDBIP}:10080/dxf/schedule?action=pause_scale_in"
curl -X POST "http://{TiDBIP}:10080/dxf/schedule?action=pause_scale_in&ttl=5h"
curl -X POST "http://{TiDBIP}:10080/dxf/schedule?action=resume_scale_in"

Parameters:

  • action: Required. Valid values are pause_scale_in and resume_scale_in.
  • ttl: Optional for pause_scale_in. Uses Go duration syntax such as 10m, 1h, or 2h45m. The default is 1h. Ignored by resume_scale_in.

Example response for pause_scale_in:

json
{
 "enabled": true,
 "ttl": 18000000000000,
 "expire_time": "2025-08-22T21:13:01.276282+08:00"
}

Example response for resume_scale_in:

json
{
 "expire_time": "0001-01-01T00:00:00Z"
}

Get or update DXF schedule tuning factors

This API gets or updates the resource tuning factors for a target keyspace. The scheduler uses these factors when calculating resource-related parameters for new DXF tasks such as IMPORT INTO and distributed add index tasks. Updating amplify_factor can make the scheduler reserve more resources for newly planned work in that keyspace.

Usage:

shell
curl "http://{TiDBIP}:10080/dxf/schedule/tune?keyspace={keyspace}"
curl -X POST "http://{TiDBIP}:10080/dxf/schedule/tune?keyspace={keyspace}&amplify_factor={number}&ttl={duration}"

Parameters:

  • keyspace: Required for both GET and POST. The target keyspace whose DXF resource calculation should be read or tuned.
  • amplify_factor: Required for POST. A floating-point number in the range [1.0, 10.0].
  • ttl: Optional for POST. Uses Go duration syntax such as 10m, 1h, or 2h45m. The default is 1h. After the TTL expires, GET returns the default tuning factor.

Example GET response:

json
{
 "amplify_factor": 1
}

Example POST response:

json
{
 "ttl": 7200000000000,
 "expire_time": "2025-09-15T22:36:24.591155+08:00",
 "amplify_factor": 2
}

Get active DXF task counts

This API returns the number of active DXF tasks in mysql.tidb_global_task, grouped by keyspace. It is useful before maintenance or upgrades because it shows tasks that have not yet finished or moved to history. This differs from /dxf/schedule/status, which only counts tasks currently considered schedulable by the scheduler.

Usage:

shell
curl http://{TiDBIP}:10080/dxf/task/active

Parameters: none.

Example response:

json
{
 "total": 3,
 "per_keyspace": {
  "SYSTEM": 1,
  "keyspace1": 2
 }
}

List DXF history tasks

This API lists rows from mysql.tidb_global_task_history in descending task ID order. Use it to inspect completed, reverted, or failed DXF tasks across keyspaces. The API uses keyset pagination: pass the previous response's NextPageToken as the next request's page_token.

Usage:

shell
curl "http://{TiDBIP}:10080/dxf/task/history"
curl "http://{TiDBIP}:10080/dxf/task/history?keyspace={keyspace}&page_size=50"
curl "http://{TiDBIP}:10080/dxf/task/history?keyspace={keyspace}&page_size=50&page_token={NextPageToken}"

Parameters:

  • keyspace: Optional. If set, only returns history tasks from that keyspace.
  • page_size: Optional. Default is 20. Valid range is [1, 200].
  • page_token: Optional. Must be a positive task ID returned as NextPageToken by the previous page.

Example response:

json
{
 "Items": [
  {
   "ID": 30001,
   "Key": "keyspace1/ddl/backfill/9",
   "Type": "backfill",
   "State": "succeed",
   "Step": -2,
   "Priority": 512,
   "RequiredSlots": 1,
   "TargetScope": "dxf_service",
   "CreateTime": "2026-04-13T11:30:10+08:00",
   "MaxNodeCount": 1,
   "ExtraParams": {},
   "Keyspace": "keyspace1",
   "StartTime": "2026-04-13T11:30:13+08:00",
   "StateUpdateTime": "2026-04-13T11:30:17+08:00",
   "EndTime": "2026-04-13T11:30:16+08:00"
  }
 ],
 "HasMore": false,
 "NextPageToken": 0,
 "ApproxTotalCount": 1
}

Get IMPORT INTO history job details

This API returns detailed history for one completed IMPORT INTO job in a target keyspace. It reads DXF history tables and returns task status, resource settings, file and KV sizes, speeds, row counts, and per-step durations. It returns 404 Not Found if the matching history task is not found.

Usage:

shell
curl http://{TiDBIP}:10080/dxf/import-into/history/job/{keyspace}/{job_id}

Parameters:

  • keyspace: Required path parameter. The keyspace where the IMPORT INTO job ran.
  • job_id: Required path parameter. The positive integer IMPORT INTO job ID.

Example response:

json
{
 "job_id": 2,
 "keyspace": "SYSTEM",
 "task_id": 2,
 "state": "succeed",
 "concurrency": 1,
 "max_node_count": 1,
 "distsql_scan_concurrency": 15,
 "index_count": 4,
 "column_count": 12,
 "file_size": "241.4MiB",
 "data_kv_size": "249.4MiB",
 "index_kv_size": "40.05MiB",
 "per_core_speed": "24.97GiB/core/hour",
 "overall_speed": "24.97GiB/hour",
 "row_count": 250000,
 "row_length": 1013,
 "duration": {
  "total": "34s",
  "encode": "2s",
  "merge_sort": "",
  "ingest": "25s",
  "collect_conflicts": "",
  "resolve_conflicts": "",
  "post_process": "4s"
 }
}

Get or update the DXF max concurrent task limit

This API gets or updates the process-local maximum number of DXF tasks that can be scheduled concurrently. The value is kept in memory only, is not persisted to TiKV, and is reset when the TiDB process restarts. When updating the value, send the request to the current DXF owner.

Get the current value:

shell
curl http://{TiDBIP}:10080/dxf/schedule/max_concurrent_task

Example response:

json
{
 "max_concurrent_task": 16,
 "persistence": "memory_only"
}

Update the value:

shell
curl -X POST "http://{TiDBIP}:10080/dxf/schedule/max_concurrent_task?value={number}"

value must be an integer in the range [16, 1000]. The response uses the same format as the GET request:

json
{
 "max_concurrent_task": 128,
 "persistence": "memory_only"
}

Update a DXF task's max runtime slots

This API updates the MaxRuntimeSlots extra parameter for one DXF task. Use it as an emergency task-level throttle, for example when a task step causes TiDB memory pressure or repeated restarts. The new value takes effect when the task executor reads the updated task metadata, commonly after the related TiDB node restarts or the task step is retried.

Usage:

shell
curl -X POST "http://{TiDBIP}:10080/dxf/task/{taskID}/max_runtime_slots?value={number}"
curl -X POST "http://{TiDBIP}:10080/dxf/task/{taskID}/max_runtime_slots?value={number}&target_step={step}&target_step={step}"

Parameters:

  • taskID: Required path parameter. The positive integer DXF task ID.
  • value: Required. A positive integer lower than the task's required_slots.
  • target_step: Optional and repeatable. A numeric business step for the task type. If omitted, the limit applies to all business steps of the task.

For IMPORT INTO tasks, common target_step values are:

ValueStep
1import
2post-process
3encode
4merge-sort
5ingest
6collect-conflicts
7conflict-resolution

For distributed add index backfill tasks, common target_step values are:

ValueStep
1read-index
2merge-sort
3ingest
4merge-temp-index

Example response:

json
{
 "task_id": 2,
 "task_key": "SYSTEM/ImportInto/2",
 "required_slots": 7,
 "max_runtime_slots": 5,
 "target_steps": [
  "post-process",
  "encode"
 ]
}