![]() |
![]() |
[ Bottom of Page | Previous Page | Next Page | Contents | Index ]
Determine in more detail the objectives of what you must monitor and what attributes are required to meet the objectives. Many of these values are default values that you can change later in the IBM Tivoli Monitoring profiles. Whenever possible, use the default values.
In this example, you can expand upon your overall objective of optimizing CPU usage by monitoring which processes are monopolizing resources. You can also examine why and when these processes require a large percentage of CPU.
Define how often you want the resource model to cycle through the resources to collect data. Minimize the impact of the monitoring activity on system performance by selecting an appropriate cycle time. Consider two main factors: how fast a given resource changes status, and how the cycling increases overhead on the system. In this example, the resources change their status quite frequently. A cycle time of 60 seconds is a reasonable compromise between monitoring status variations and saving system performance.
In this example, you do not need to examine processes with low CPU consumption because they do not affect CPU performance too severely. You can apply a filter to our data collection so that you collect data only for the processes using at least 20% of CPU capacity.
You can also sort the results in ascending order of CPU usage. You can configure our resource model to monitor only the specific processes that exceed a threshold. If none of our processes exceed the 20% CPU usage threshold, no processes are monitored. This configuration optimizes system performance during monitoring.
[ Top of Page | Previous Page | Next Page | Contents | Index ]