Database Reference
In-Depth Information
Chapter 10
Metric Extensions and Management
Plug-ins
by Alex Gorbachev
The 10g and 11g versions of Oracle Enterprise Manager introduced some basic features aimed at helping the product
become an enterprise manageability and monitoring framework. However, OEM was, above all, a manageability tool
built for Oracle products, with some features extending its capabilities to other target types not supported out of the
box. Oracle Enterprise Manager Cloud Control 12c has been completely rearchitected from the ground up to be, first
of all, a framework for enterprise manageability. Nearly all functionality for management and monitoring of specific
Oracle and third-party products has been built on top of that framework as plug-ins.
Just like previous versions of OEM, EM12c provides options for users and product vendors to extend its
capabilities in two basic ways. The first part of this chapter presents the simplest: metric extensions. The second part
covers the basics of management plug-in development as well as some advanced features.
I recommend that you try the hands-on examples as you read through this chapter, and have the EM12c
extensibility documentation handy as a reference. Full topic coverage would require a book of its own, and this
chapter is more of an introduction to extending EM12c.
However, don't skip this chapter even if you don't expect to develop plug-ins yourself. Knowledge of EM12c
internals will help you better understand how EM12c works. Ultimately, this will enable you to become more
productive and find the environment more intuitive to use.
Metric Extensions
Metric extensions in EM12c replace user-defined metrics (UDMs) that were initially introduced in OEM 10g and are
now deprecated in 12c. Compared to UDMs, metric extensions provide much more flexibility in data collection and
are easier to deploy and manage at scale.
As you should already know by now, each monitored target comes with a set of predefined metrics that are
designed by the plug-in creator (which could be Oracle or a third-party vendor). For example, host targets collect CPU
utilization metrics, and Oracle Database targets collect physical IO metrics, among others. Usually these metrics will
be enough, but at times you might want to collect more metrics than are available out of the box in order to monitor
conditions specific to your environment or applications.
Metric extensions are run by an agent from the host that the agent is deployed on. While you are creating and
managing metric extensions centrally by using the OMS, they need to be distributed to the agents that monitor targets
where metric extensions are deployed to. That's done automatically as soon as a metric extension is deployed on the
first target of an agent.
 
Search WWH ::




Custom Search