跳到主要内容
版本：Nightly

Admin API

Admin 提供了一种简单的方法来查看集群信息，包括 metasrv 健康检测、metasrv leader 查询、数据库元数据查询和数据节点心跳检测。

Admin API 是一个 HTTP 服务，提供一组可以通过 HTTP 请求调用的 RESTful API。Admin API 简单、用户友好且安全。 可用的 API：

  • /health
  • /leader
  • /heartbeat
  • /maintenance

所有这些 API 都在父资源 /admin 下。

在以下部分中，我们假设你的 metasrv 实例运行在本地主机的 4000 端口。

/health HTTP 端点​

/health 端点接受 GET HTTP 请求，你可以使用此端点检查你的 metasrv 实例的健康状况。

定义​

curl -X GET http://localhost:4000/admin/health

示例​

请求​

curl -X GET http://localhost:4000/admin/health

响应​

OK

/leader HTTP 端点​

/leader 端点接受 GET HTTP 请求，你可以使用此端点查询你的 metasrv 实例的 leader 地址。

定义​

curl -X GET http://localhost:4000/admin/leader

示例​

请求​

curl -X GET http://localhost:4000/admin/leader

响应​

127.0.0.1:4000

/heartbeat HTTP 端点​

/heartbeat 端点接受 GET HTTP 请求，你可以使用此端点查询所有数据节点的心跳。

你还可以查询指定 addr 的数据节点的心跳数据，但在路径中指定 addr 是可选的。

定义​

curl -X GET http://localhost:4000/admin/heartbeat
查询字符串参数类型可选/必选定义
addrString可选数据节点的地址。

示例​

请求​

curl -X GET 'http://localhost:4000/admin/heartbeat?addr=127.0.0.1:4100'

响应​

[
 [{
  "timestamp_millis": 1677049348651,
  "cluster_id": 0,
  "id": 1,
  "addr": "127.0.0.1:4100",
  "is_leader": false,
  "rcus": 0,
  "wcus": 0,
  "table_num": 0,
  "region_num": 2,
  "cpu_usage": 0.0,
  "load": 0.0,
  "read_io_rate": 0.0,
  "write_io_rate": 0.0,
  "region_stats": []
 }, {
  "timestamp_millis": 1677049344048,
  "cluster_id": 0,
  "id": 1,
  "addr": "0.0.0.0:4100",
  "is_leader": false,
  "rcus": 0,
  "wcus": 0,
  "table_num": 0,
  "region_num": 2,
  "cpu_usage": 0.0,
  "load": 0.0,
  "read_io_rate": 0.0,
  "write_io_rate": 0.0,
  "region_stats": []
 }, {
  "timestamp_millis": 1677049343624,
  "cluster_id": 0,
  "id": 1,
  "addr": "127.0.0.1:4100",
  "is_leader": false,
  "rcus": 0,
  "wcus": 0,
  "table_num": 0,
  "region_num": 2,
  "cpu_usage": 0.0,
  "load": 0.0,
  "read_io_rate": 0.0,
  "write_io_rate": 0.0,
  "region_stats": []
 }, {
  "timestamp_millis": 1677049339036,
  "cluster_id": 0,
  "id": 1,
  "addr": "0.0.0.0:4100",
  "is_leader": false,
  "rcus": 0,
  "wcus": 0,
  "table_num": 0,
  "region_num": 2,
  "cpu_usage": 0.0,
  "load": 0.0,
  "read_io_rate": 0.0,
  "write_io_rate": 0.0,
  "region_stats": []
 }, {
  "timestamp_millis": 1677049338609,
  "cluster_id": 0,
  "id": 1,
  "addr": "127.0.0.1:4100",
  "is_leader": false,
  "rcus": 0,
  "wcus": 0,
  "table_num": 0,
  "region_num": 2,
  "cpu_usage": 0.0,
  "load": 0.0,
  "read_io_rate": 0.0,
  "write_io_rate": 0.0,
  "region_stats": []
 }, {
  "timestamp_millis": 1677049334019,
  "cluster_id": 0,
  "id": 1,
  "addr": "0.0.0.0:4100",
  "is_leader": false,
  "rcus": 0,
  "wcus": 0,
  "table_num": 0,
  "region_num": 2,
  "cpu_usage": 0.0,
  "load": 0.0,
  "read_io_rate": 0.0,
  "write_io_rate": 0.0,
  "region_stats": []
 }, {
  "timestamp_millis": 1677049333592,
  "cluster_id": 0,
  "id": 1,
  "addr": "127.0.0.1:4100",
  "is_leader": false,
  "rcus": 0,
  "wcus": 0,
  "table_num": 0,
  "region_num": 2,
  "cpu_usage": 0.0,
  "load": 0.0,
  "read_io_rate": 0.0,
  "write_io_rate": 0.0,
  "region_stats": []
 }, {
  "timestamp_millis": 1677049329002,
  "cluster_id": 0,
  "id": 1,
  "addr": "0.0.0.0:4100",
  "is_leader": false,
  "rcus": 0,
  "wcus": 0,
  "table_num": 0,
  "region_num": 2,
  "cpu_usage": 0.0,
  "load": 0.0,
  "read_io_rate": 0.0,
  "write_io_rate": 0.0,
  "region_stats": []
 }, {
  "timestamp_millis": 1677049328573,
  "cluster_id": 0,
  "id": 1,
  "addr": "127.0.0.1:4100",
  "is_leader": false,
  "rcus": 0,
  "wcus": 0,
  "table_num": 0,
  "region_num": 2,
  "cpu_usage": 0.0,
  "load": 0.0,
  "read_io_rate": 0.0,
  "write_io_rate": 0.0,
  "region_stats": []
 }, {
  "timestamp_millis": 1677049323986,
  "cluster_id": 0,
  "id": 1,
  "addr": "0.0.0.0:4100",
  "is_leader": false,
  "rcus": 0,
  "wcus": 0,
  "table_num": 0,
  "region_num": 2,
  "cpu_usage": 0.0,
  "load": 0.0,
  "read_io_rate": 0.0,
  "write_io_rate": 0.0,
  "region_stats": []
 }]
]

/maintenance HTTP 端点​

集群维护模式是 GreptimeDB 中的一项安全功能，它可以临时禁用自动集群管理操作。此模式在集群升级、计划停机以及任何可能暂时影响集群稳定性的操作期间特别有用。有关更多详细信息，请参阅集群维护模式。

/procedure-manager HTTP 端点​

该端点用于管理 Procedure Manager 状态。有关更多详细信息，请参阅防止元数据变更。