Microsoft® Windows® 2000 Scripting Guide
Many of the important operations that take place on a computer (especially on servers) run as services. This makes it imperative that you carefully monitor the services running on the computers in your network. Many services (such as DNS and DHCP) are so critical that a failure on a single server could adversely affect hundreds or even thousands of users, preventing them from logging on to the network or accessing network-based resources.
In general, there are three forms of service monitoring:
Monitoring service availability. Measures the percentage of time that a service is available.
The exact definition of availability depends on the expectations for each service. If a database service must be available to users from 8:00 A.M. to 6:00 P.M. Monday through Friday, it can be considered completely available as long as the service is running during those times. If the service fails on a Saturday, or at 2:00 A.M. on Tuesday, this does not affect availability. It does, however, affect service reliability.
Monitoring service reliability. Measures how frequently a service fails and the amount of time required to restore a failed service to full functionality.
Reliability is calculated by dividing the time the service is functioning by the total number of days in a year. For example, a service that experiences a total downtime of 2 days during the course of a year is 99.5 percent reliable (363 days of availability divided by 365 days in a year).
Monitoring service performance. Measures whether the service carries out its tasks in the expected manner (for example, whether the service handles the expected number of requests in the expected amount of time). Although it is possible to use WMI to monitor the performance of some operating system services in Windows 2000, a complete discussion of this type of monitoring is beyond the scope of this chapter.