Predicting Oracle Performance Issues

Database performance remains to be a critical in ensuring application performance. Reacting to database performance problems as they occur, however, is an increasingly ineffective strategy for maintaining database performance; SQL optimization technology has advanced so far that "trivial" SQL issues—those easily fixed by reactively tuning SQL statements—are rare. The remaining problems require non-trivial solutions, which typically involve significant disruption to production systems and cannot be implemented without significant lead time. 

This paper discusses two main categories of Oracle performance problems that we want to predict:

1. SQL statements that are degrading (increasing in response time and resource consumption) and that will, if not tuned, eventually overload the system and cause service-level issues.

2. Database bottlenecks that occur when demand on the database exceeds configured capacity. These bottlenecks can manifest themselves as lock contention, I/O degradation and so on.

See how you can detect these performance problems ahead of time, ensuring business continuity and reducing the cost of database-driven performance "brown outs".