Interactive tools operate in real time. When you log in to Oracle using SQL*Plus and run a script, you are being interactive. Interactive tools have the distinct advantage of going technically very deep. All the data is available because the activity is happening now. Hyper-focused interactive tools usually have dazzlingly graphics. However, be aware that interactive tools do have limitations. During a firefight, there is usually a lot of chaos, and while you are focusing on one problem, another problem is occurring. If you are limited to interactive tools, you could miss a key investigative opportunity.

Historically focused tools periodically collect data and then allow you to "look back in time," as they say. Examples of historical tools are Statspack and AWR. Historical tools are wonderful for analyzing a problem after it happened (like while you slept last night), when multiple problems need to be analyzed, or if you need to reference a past situation (this occurs more often than people think). Historical tools usually provide trending and graphing capabilities, which are especially helpful in communicating with nontechnical people.

Historical product vendors have a very technically challenging problem. They want to gather as much data as possible and as frequently as possible. However, vendors know the more frequently they gather data and the more data they gather, the more their collection activities will impact the system. This is one reason why good historical products seem expensive. In Chapter 5, which covers Oracle performance diagnosis, I will address data-collection challenges.