 |
|
|
Metric dba_hist Views The metric dba_hist data dictionary views are organized into several groups: system; session; service; file; tablespace; and event metrics. All available metric groups can be found in the v$metricgroup dynamic view:
SQL> select * from V$METRICGROUP;
GROUP_ID NAME INTERVAL_SIZE MAX_INTERVAL -------- -------------------------------- ------------- ------------ 0 Event Metrics 6000 1 1 Event Class Metrics 6000 60 2 System Metrics Long Duration 6000 60 3 System Metrics Short Duration 1500 12 4 Session Metrics Long Duration 6000 60 5 Session Metrics Short Duration 1500 1 6 Service Metrics 6000 60 7 File Metrics Long Duration 60000 6 9 Tablespace Metrics Long Duration 6000 0
All available metrics for which the AWR keeps history are in the dba_hist_metric_name view that stores snapshots for the v$metric_name view. There are approximately 184 different metrics available.
Metrics are actively used by sophisticated automation features of Oracle10g such as the Automatic Memory Management (AMM) and advisory engines such as Automatic Database Diagnostic Monitor (ADDM), SQL Advisor , etc. For example, the ADDM uses a goal-based algorithm designed to minimize db time . This metric is computed as a cumulative time spent.
SQL> select METRIC_NAME from DBA_HIST_METRIC_NAME;
The following is a list of several metrics that are computed by the AWR snapshot engine; however, the list is not exhaustive.
METRIC_NAME -------------------------------------- Average File Read Time (Files-Long) Average File Write Time (Files-Long) Average Users Waiting Counts Background Checkpoints Per Sec Blocked User Session Count Branch Node Splits Per Sec Branch Node Splits Per Txn Buffer Cache Hit Ratio CPU Time (Session) CPU Time Per User Call CPU Usage Per Sec
The usage of metric statistics in time-series tuning approach will be presented later in this book. This chapter provides a general overview of metric dba_hist views that are available and a short description of information they provide.
|
|
|
