Correlation Analysis Reports with AWR and ASH Views

This example should demonstrate how the AWR and ASH data can be used to create an almost infinite number of sophisticated custom performance reports.

AWR and ADDM in a Nutshell

Using the wealth of metrics within the AWR can be greatly useful for the senior DBA who wants to obtain detailed correlation information between any of the 500+ performance metrics captured by the AWR.

The AWR repository can also be used for Oracle Data Mining.  Using The Oracle Database 10g Data Mining (ODM) option, the Oracle DBA can automatically scan the AWR seeking statistically significant correlations between metrics using multivariate Chi-Square techniques to reveal hidden patterns within the performance information.  The Oracle Database 10g ODM uses sophisticated Support Vector Machines (SVM) algorithms for binary, multi-class classification models, and has built-in linear regression functionality.

As Oracle Database 10g continues to evolve, Oracle will continue to enhance the mechanisms for analyzing the valuable performance information in AWR.  At the present rate, future releases of Oracle may have true artificial intelligence built-in to detect and correct even the most challenging Oracle optimization issues.

The creation of AWR and ASH provides a complete repository for diagnosing and fixing any Oracle performance issue.

The AWR dba_hist views are similar to the well-known STATSPACK tables, making it easy to migrate existing performance reports to Oracle Database 10g. The dba_hist views are fully-documented and easy to use for writing custom scripts.