The preceding trace file snippet is what we would expect, with one exception. It does contain the standard trace information about CPU consumption and block activity, and it also contains wait event data including the average wait times. As discussed previously, IO-related average wait times are very useful in understanding Oracle's interaction with the IO subsystem. What is missing though is bind variable details. The SQL statement clearly shows bind variables are being used (and they are system-generated), but notice there are no references to the actual bind variables. When multiple trace files are consolidated, don't expect to see the individual bind variables. To get this level of detail, you'll need to look at the trace file yourself or use another tool.

DBMS_MONITOR is a wonderful Oracle package. While it does not solve all profiling challenges, it is truly a giant leap forward. It's flexibility in purpose and use, and the fact that no additional license is required make it something every Oracle DBA should at least be familiar with.

Some would say instrumentation is the best performance statistics gathering approach, but others would say periodic sampling is the best. After all, nearly all third-party tools rely on polling, which is sampling. Just as Iron Man's Tony Stark would say, "Is it too much to ask for both?" And this is exactly what Oracle has done.