Best Practices and Pitfalls

The best way to approach management is to use early planning and design applications with management in mind. The application design phase is typically the best time to determine, using logging instrumentation, how to disable, enable, minimize or maximize information output. Another key component to successful management is to ensure that systems can be managed accordingly at all stages of the development cycle including QA and production. Production-level management is essential for monitoring the overall IT environment. Remote monitoring and alert notification systems can be utilized to enhance control of the distributed heterogeneous applications. Additionally, individual management techniques should be utilized to monitor the hardware, operating system, infrastructure, and finally application layer. Being able to monitor systems' uptime and quickly troubleshoot management problems by drilling down to the corresponding layer to determine the failing module is an important factor for success. Implementing comprehensive management as discussed by the ISO characteristics has to be adequately addressed. Monitoring alone would not be sufficient if the fault management is not addressed. As an example, a failing instance of an application server needs to be transparently replaced by a working instance.

Common pitfalls in the management space include excessive logging of applications deployed in a production environment, causing performance degradation. Oppositely, lack of application instrumentation can affect the amount of time the problem takes to troubleshoot. Lack of discrete layer separation across various managed layers also degrades quality of management. Situations where an application flaw, an infrastructure component such as application server, or an OS unknowingly causes a problem require diverse skill sets and are typically time consuming. Therefore for successful IT management, it is critical to address management at individual layers.