Custom Script(s)

This free method to monitoring Oracle GoldenGate consists of using a script that will be run on each server running Oracle GoldenGate.  The script needs to issue “ggsci” commands and then parse the resulting output looking for issues (evaluating process status, lag and checkpoint time, etc).   Of course, this script has to be written and the logic ascertained as to how to monitor GoldenGate.  You do have the advantage of using any language you prefer for your platforms involved and can configure the monitoring logic however you like as well. 

This is a very popular method for monitoring Oracle GoldenGate, as there are many advantages and it is “free” from product licensing costs.  To be brutally honest, this method has become very very popular due to the many issues when using the Oracle provided products (OEM Plug-in, Monitor and Director); those tools are indeed getting better and better, but the needed fixes and features where not in the products at the time needed by many GoldenGaters implementing in production environments.  Certainly there are times where a robust, cheap, easily customizable and an existing and deployable solution is needed, as well unlimited integration with existing enterprise monitoring tools already deployed,  which may not be OEM. 

GoldenGate is a great product with many uses, but monitoring must be robust and usable across all the GoldenGate instances/versions out in the wild.  Therefore, the custom script approach in conjunction with the heartbeat is many times the only solution used and it can cover the required monitoring aspects.

As part of this Blog, we are providing/sharing a free script for you to use in your quest to monitor Oracle GoldenGate. The script we are offering has been written in Perl for portability (Windows, Linux/UNIX and AS400) and performance reasons, and has been deployed to hundreds of instances of GoldenGate across many different environments/platforms. Certainly the script may not do all you want to do, but it is a great start and robust at monitoring which is needed now and across all platforms, configurations and versions.

The script and documentation can be found here in separate pages!  Script is called "gg_monitor.pl" and the documentation that I have so far for it.

Warning, the documentation is a little light (currently as write), but I have tried to cover it well enough to understand what is needed and how to set it up and use it/modify it easily to suite your needs.

The script can be run through almost any method capable of running scripts (ie cron, through a database job that runs a shell script, or via third-party schedulers or monitoring tools such as OEM, OpenView, Tivoli, Big Brother, Nagios, etc, etc that can call a script to monitor an application).

This script can also be configured to run as part of cluster scripts for a clustered solution. It is not uncommon to find Oracle GoldenGate registered as a cluster resource on an Oracle RAC (for failover between nodes in the RAC). When node failure occurs, someone has to change the script (or monitoring application) to monitor GoldenGate on the new node. Therefore, this script can be registered as a CRS resource (with a small bit of editing) and when GoldenGate fails over to another node, the script fails over with it as well and starts monitoring OGG wherever it is running from (on that new node). This is a monitoring life saver for many clustered GoldenGate implementations.

Scripts (or this script) can also be used as a metric extension in OEM as well, a fancy way of using a third-party scheduler such as OEM. However, running any script as a metric extension in OEM requires OEM and OEM agents deployed. If your enterprise does not use OEM, then the script approach is still an excellent solution.