common infor need to get from Statspack and AWR

1.ratio of Logical Reads to Physical Reads

  The ratio should be 100s or 1000s : 1. 

2.if redo size per second is higher than usual,it implies update or insert and will obviously take more time and effort 

3.poor Database cache hit rate leads to high physical I/O.

4.For Buffer Hit, much higher much better,near to 100% will be the best.

 V$SYSMETRIC

5.Logical Reads per second and the ratio as compared to the redo size

This ratio has an averaged ~ 10:1 for the whole hour period  indicating very high read activity ,this is likely to cause high CPU consumption

6.(latch: cache buffers chains) generally point to same block
being accessed many times

7.The reason for db file scattered read

 SQL Tuning issues, lack of indexes, IO contention, etc

8.Response Time = Service Time + Wait Time

'Service Time' is measured using the statistic 'CPU used by this session' 

'Wait Time' is measured by summing up time spent on Wait Events

 we need to calculate percentages for all the 'Response Time' components ,and then we can detimine if the event in the top 5 event is the bottleneck

Reference note 565812.1 276103.1