从 ASH 获取最近几分钟的 Top 等待事件和 Top SQL

命令

-- Last 30 min: wait event breakdown (each sample ~1 s of active session time)
SELECT NVL(event, 'ON CPU') AS event, wait_class,
       COUNT(*) AS samples,
       ROUND(RATIO_TO_REPORT(COUNT(*)) OVER () * 100, 1) AS pct
FROM   v$active_session_history
WHERE  sample_time > SYSTIMESTAMP - INTERVAL '30' MINUTE
GROUP  BY NVL(event, 'ON CPU'), wait_class
ORDER  BY samples DESC
FETCH FIRST 10 ROWS ONLY;

-- Last 30 min: top SQL (with CPU / User I/O breakdown)
SELECT sql_id, sql_opname,
       COUNT(*) AS samples,
       SUM(CASE WHEN session_state = 'ON CPU' THEN 1 ELSE 0 END) AS cpu,
       SUM(CASE WHEN wait_class = 'User I/O' THEN 1 ELSE 0 END)  AS user_io,
       ROUND(RATIO_TO_REPORT(COUNT(*)) OVER () * 100, 1) AS pct
FROM   v$active_session_history
WHERE  sample_time > SYSTIMESTAMP - INTERVAL '30' MINUTE
  AND  sql_id IS NOT NULL
GROUP  BY sql_id, sql_opname
ORDER  BY samples DESC
FETCH FIRST 10 ROWS ONLY;

-- Older periods: DBA_HIST_ACTIVE_SESS_HISTORY (one sample every 10 s; COUNT(*)*10 = seconds)

-- Ready-made ASH report (last 30 min):
SET LINESIZE 8000 PAGESIZE 0 TRIMSPOOL ON HEADING OFF FEEDBACK OFF
SELECT output
FROM   TABLE(DBMS_WORKLOAD_REPOSITORY.ASH_REPORT_TEXT(
         l_dbid     => (SELECT dbid FROM v$database),
         l_inst_num => (SELECT instance_number FROM v$instance),
         l_btime    => SYSDATE - 30/1440,
         l_etime    => SYSDATE));
-- Interactive: @?/rdbms/admin/ashrpt.sql

作用

使用 Active Session History 样本，以百分比显示最近几分钟内哪些等待事件和 SQL 语句消耗了 DB time；无需等待 AWR 快照即可诊断事件。

适用场景

实时性能下降期间、在 AWR 平均值中消失的短时尖峰，以及查看特定会话或 SQL 在等待什么时。

注意

查询 V$ACTIVE_SESSION_HISTORY 和 DBA_HIST_ACTIVE_SESS_HISTORY 需要 Diagnostics Pack 许可。ASH 是采样数据：样本数用于估算时间，并非精确测量，极短的 SQL 可能不会出现。在繁忙系统上，内存中的 ASH 缓冲区只能保存几个小时。