| Lesson 10 || Customizing Oracle alerts |
| Objective || Identify database events to be monitored by the OEM console. |
Customizing Oracle Alerts
Once Tcl has been mastered, you can write customized events that will be monitored by the OEM console. Generally, you will want alerts for any
database event that may impact the availability of your databases. For example:
- A tablespace is greater than 95 percent full.
- A table is approaching maximum extents.
- The archived redo log filesystem is greater than 95 percent full.
- A table is unable to take a new extent.
- An Oracle listener is not running or cannot accept connections.
- ORA-600 errors are appearing in the Oracle alert log.
- The buffer hit ratio has fallen below 70 percent.
As you can see, many of these alerts are proactive, because the database has not yet experienced a critical problem.
With expedient notification, the DBA will be able to intercede and correct the problem before the database becomes unavailable.
Of course, there is no one-size-fits-all approach to event monitoring and exceptions need to be addressed. For example, you cannot globally monitor:
- Tablespaces greater than 95 percent full, since read-only tablespaces may be packed with data to 100 percent
- The Oracle rollback segments, which commonly reach maximum extents and do not warrant an alert
- Databases in NOARCHIVELOG mode, which do not need a monitor for an archived redo log file system In short, exceptions need to be addressed for almost every remote database.
These exceptions will need to be coded into the Tcl monitors.
Managing Oracle Database
Oracle includes many features that make the database easier to manage.
Oracle management fundamentally improved with the introduction of Oracle Database 10g and has continued to evolve toward being more self-tuning and self-managing in subsequent database releases.
If you are still managing Oracle Databases using older techniques (for example, scripts), you might want to reevaluate your thinking on management. Since Oracle Database 10g, statistics are automatically gathered to an Automatic Workload Repository (AWR) within the database.
Oracle’s (ADDM) Automatic Database Diagnostic Monitor evaluates the statistics on a regular basis and sends alerts of potential problem conditions to Oracle Enterprise Manager, where you can evaluate the condition in more detail and potentially take corrective actions.
Some of the newer fully automated features, such as Automatic Memory Management, also leverage data gathered in the AWR.
Oracle has a near real-time view of current database conditions as it makes automated recommendations. Such recommendations will often be more accurate than would be possible with the manual processes you might have used in the past. In the following subsections, we will introduce the impact this has on Oracle Enterprise Manager and addon packs, Information Lifecycle Management, backup and recovery, and database availability.
The next lesson concludes this module.