监控可用性组健康状况与延迟（log send / redo 队列、RPO）

命令

-- Replica health (run on the primary to see all replicas):
SELECT ag.name AS ag_name, ar.replica_server_name, ars.role_desc,
       ar.availability_mode_desc, ar.failover_mode_desc,
       ars.connected_state_desc, ars.synchronization_health_desc,
       ars.operational_state_desc, ars.last_connect_error_description
FROM   sys.availability_groups ag
  JOIN sys.availability_replicas ar                ON ar.group_id = ag.group_id
  JOIN sys.dm_hadr_availability_replica_states ars ON ars.replica_id = ar.replica_id
ORDER BY ag.name, ars.role_desc;

-- Per-database lag: log send queue (RPO), redo queue (RTO):
SELECT ag.name AS ag_name, ar.replica_server_name, DB_NAME(drs.database_id) AS db,
       drs.is_primary_replica, drs.synchronization_state_desc, drs.synchronization_health_desc,
       drs.is_suspended, drs.suspend_reason_desc,
       drs.log_send_queue_size  AS log_send_queue_kb,
       drs.log_send_rate        AS log_send_kb_s,
       drs.redo_queue_size      AS redo_queue_kb,
       drs.redo_rate            AS redo_kb_s,
       CAST(drs.redo_queue_size * 1.0 / NULLIF(drs.redo_rate, 0) AS DECIMAL(18,1)) AS est_redo_catchup_s,
       drs.last_commit_time,
       DATEDIFF(SECOND, drs.last_commit_time, pri.last_commit_time) AS data_loss_window_s,
       drs.secondary_lag_seconds
FROM   sys.dm_hadr_database_replica_states drs
  JOIN sys.availability_replicas ar ON ar.replica_id = drs.replica_id
  JOIN sys.availability_groups ag   ON ag.group_id = drs.group_id
  LEFT JOIN sys.dm_hadr_database_replica_states pri
         ON pri.group_database_id = drs.group_database_id AND pri.is_primary_replica = 1
ORDER BY ag.name, db, drs.is_primary_replica DESC;

-- WSFC quorum state:
SELECT cluster_name, quorum_type_desc, quorum_state_desc FROM sys.dm_hadr_cluster;
SELECT member_name, member_type_desc, member_state_desc, number_of_quorum_votes
FROM   sys.dm_hadr_cluster_members;

作用

显示副本连接和同步健康状况；按数据库给出 log send 队列、redo 队列、发送/redo 速率、预计 redo 追平时间以及与主副本的最后提交时间差（潜在数据丢失窗口）；还返回 WSFC 仲裁状态。

适用场景

AG 仪表板出现 NOT_HEALTHY / NOT SYNCHRONIZING 告警、计划内故障转移前检查 RPO/RTO，或用户反映可读辅助副本上数据过时时。

注意

不断增长的 log_send_queue 是主副本故障时会丢失的日志量（RPO）；不断增长的 redo_queue 决定故障转移后数据库联机所需的时间（RTO）。在辅助副本上只能看到本地副本；在主副本上运行可看到全貌。data_loss_window_s 仅在有写入流量时有意义。只读查询；安全。