The Performance Sentry collection service gathers performance Objects and Counters from the Windows Server operating system and writes them to the NTSMF collection file. Using Performance Sentry, you can collect any of the performance Objects that are available on your machine, including performance statistics on resource usage, networking, and application performance. In addition, Performance Sentry collects static configuration information about your Windows 2000 machine, which allows you to inventory your hardware and software assets.
In a production environment, Performance Sentry is designed to start automatically and run unattended. The Collection service collects performance data for just the Objects and Counters you specify at regularly scheduled intervals. Following installation, the Performance Sentry Collection service starts automatically and collects data continuously.
There is a substantial amount of performance measurement data that is potentially available on typical Windows servers and workstations. It is neither necessary or desirable to collect all the data that is available all the time. The Performance Sentry collection service is an intelligent agent that is designed to capture only the most important metrics to support problem detection and resolution. Collection sets tailored to your environment can be defined and assigned to the collection service. An intelligent filtering mechanism restricts the number of entries in the data collection file according to filtering criteria that you specify. The reduced size of the data collection file that results means less transfer time across the network and less processing time at its final destination.
The Collection service is designed to execute continuously in an unattended mode. It creates data collection files automatically, generating unique names of the form systemname.yyyymmddhhmmcustomqualifier.smf. At regular intervals using our own scheduling facilities, the Collection service closes the current data collection file and opens a new one. The older data collection file becomes immediately available for processing. This process is repeated at the end of each collection cycle. You specify when a collection cycle begins and the duration of each cycle. You also specify the processing you want performed (such as transferring the collection file to a shared folder for consolidation and further processing) at each collection cycle end.
At the end of each collection cycle, you will typically summarize and copy the .smf data collection file to some other location where it is consolidated with other collection files and processed to update a performance database that is used for reporting and capacity planning. After it has been created, the collection service’s file management capabilities automatically age the collection files stored on the local system and ultimately delete them, all according to your specifications. The collection service can share the active collection file with other applications that want to process data from the current collection interval. For best results, we recommend you make a copy of the active collection file before processing it using the Performance Sentry Portal or SAS.