Figure 5-6. This is an example of gathering CPU consumption for the database writer background process. Notice v$sess_time_model contains a value greater than zero, where v$sesstat always will show zero. It appears Oracle is not including the database writer's CPU consumption in the instance CPU consumption statistic. Also, by repeatedly querying v$sess_time_model, the value was observed to be updated about every 6 seconds.

Figure 5-7 contains the same SQL as in Figure 5-6, but Figure 5-7 is reporting on an active server process, whereas Figure 5-6 is looking at the database writer background process. As we would expect, the time model CPU consumption is recorded in the DB CPU statistics as 430,836 ms, which is 0.43 second. The v$sesstat CPU consumption records the server processes has consumed 39 cs, which is 0.39 second. This time difference should not surprise you. If you take any Statspack or AWR report and compare these statistics, you will notice there is always a difference-but probably not enough to throw off the v$sysstat-based service time analysis. Since the time model is based on increased instrumentation granularity, if possible, always use the time model views.

Figure 5-7. This is a sample CPU consumption query for an Oracle server process. Notice the v$sess_time_model CPU consumption is properly allocated in the DB CPU bucket. Also notice the CPU times are slightly different. The time model view shows CPU consumption at 0.43 second, whereas the v$sesstat view shows it to be 0.39 second.