Logging
One benefit of dbm_scheduler over dbms_job is the fact that it records job actions and (if needed) the actions on the jobs. This enables us to find out when jobs ran and compare the runtime behavior based on recorded execution times, instead of vague user estimations and assumptions. Not everything we would wish for is recorded. It is very useful to add some performance metrics to the detailed logging. This can help us spot where the longer job runtime came from. For example, when we see that a job that normally runs for 2 seconds and performs 4,000 buffer gets, now ran for 3 hours and performed 800,000,000 buffer gets, it's clear that some investigation is required.
The disadvantage of this logging is that it has to be configured and maintained. For this, we have the log_detail_level and the log_history parameters of the job, job_class, or the Scheduler.
Log detail level
The log detail level can be defined at several locations. It can be defined in the job creation and job_class, where...
