Lesson 2 | What to monitor? |
Objective | List terms to monitor on a regular basis. |
What to monitor
When managing an Oracle database, you should be monitoring a variety of items on a
regular basis. The table below lists the items and the reasons they need to be monitored:
Item | What to monitor |
Database alert log | This is a text file to which Oracle writes serious error messages. Oracle also generates log entries for significant events such as database startup and shutdown. |
License compliance | Most Oracle licenses specify a maximum number of users. Are you exceeding that number? |
Disk space | How much are you using currently? How much free space do you have? When will you run out of space? |
Buffer cache hit ratio | Do you have enough memory allocated to the database buffer cache? |
Dictionary cache and library cache hit ratios | Do you have enough memory allocated to the shared pool? |
File I/O statistics | Is your I/O load evenly distributed across multiple disks? |
Redo logs | Do you have enough redo log files so that users aren't forced to wait while one is archived? Are you experiencing buffer contention? |
What are Dynamic Performance Views in Oracle?
Dynamic Performance Views
Monitoring many of these items involves selecting data from Oracle's dynamic performance views. Dynamic performance views are the ones owned by the user named SYS
and have names that begin with V$. In fact, DBAs commonly refer to these as the "V$ views" (vee-dollar). Oracle implements a large number of V$ views, and the information that they return constantly changes to reflect changes in the state of the database over time.
In the next lesson, I will describe the contents of the alert log.