The Performance Monitor is a module to facilitate monitoring the performance of Adaptive Server. The Performance Monitor is based on the sophisticated build-in monitoring tables in Adaptive Server (15.0 or later), which provides database statistical and diagnostic information. You can easily understand the performance behavior of Adaptive Server by monitoring its run time statistics by means of snapshots, real time monitoring, and scheduled monitoring. The monitor result is presented graphically from the overall database performance down to the statistics of an individual process.
The Performance Monitor displays what is happening on the database. It helps to identify a CPU or I/O intensive server that may require some configuration tuning in the operating system or Adaptive Server. In addition, it also helps to illustrate that the tuning of application code and database design are where the most performance improvement can normally be gained. A missing index on a table or a poorly constructed SQL statement may cause poor performance on the database causing a chain reaction from high CPU, I/O, and Lock Waits to even more Dead Locks. The Performance Monitor can easily help to locate the symptom, but you still need the other modules for solving the performance problems, using the Index Advisor to locate a missing index or the SQL Optimizer to rewrite a poorly performing SQL statement.
The Performance Monitor monitors for the statistics that are consuming the most resources. It identifies where the top resource consumption occurs through various performance statistics and thereby filtering out unnecessary information so that performance problems are easier to pin point.
A daily monitoring of the performance of the database can be achieved by setting up a periodic capture time for retrieving the performance statistics. For example; to review the database performance on daily basis, a 15 minutes to 60 minutes refresh interval can provide a pretty clear picture of the system performance and resource consumption status; such as CPU, Devices IO, Network I/O, Lock/Dead Lock, etc. For troubleshooting on an extremely slow database, a frequent polling to the database may be too intrusive, so the manual refresh option which takes snapshots from the database may be more appropriate.