View the Exhibit to examine the output of the DBA_OUTSTANDING_ALERTS view. 
After 30 minutes, you executed the following command:
SQL> SELECT reason,metric_value FROM dba_outstanding_alerts;
REASON METRIC_VALUE
------------------------------ ------------
Tablespace [TEST] is [28 perce 28.125
nt] full
What could be the reason for the elimination of the other rows in the output?
A.An Automatic Workload Repository snapshot has been taken recently.
B.The non-threshold-based alerts are transferred to DBA_ALERT_HISTORY.
C.The threshold alerts conditions are cleared and the alerts are transferred to DBA_ALERT_HISTORY.
D.The threshold alerts related to database metrics are permanently stored in DBA_ALERT_HISTORY but not the threshold alerts related to instance metrics.
服务器生成预警分为两种类型：阈值预警和非阈值预警。
大多数服务器生成的预警是通过对数据库度量设置警告阈值和严重阈值来配置的。可为 120 多个度量定义阈值，这些度量包括：
• 每秒的物理读取次数
• 每秒的用户提交次数
•SQL 服务响应时间
除了表空间的空间占用率度量与数据库相关外，其它度量都与实例相关。阈值预警又称为“有状态预警”，清空预警条件后，会自动清空这些预警。有状态预警显示在DBA_OUTSTANDING_ALERTS 中，清空后会转到 DBA_ALERT_HISTORY。
服务器生成的其它预警对应于特定的数据库事件，如 ORA-*错误、“快照太旧”错误、“恢复区空闲空间过少”和“可恢复会话被挂起”等。这些都不是基于阈值的预警，又称为“无状态预警”。无状态预警会直接转到历史记录表。只在 Database Control 环境中，清空无状态预警才有意义，因为 Database Control 会在自身的资料档案库中存储无状态预警。