这里是否有比调整文件大小更大的目标? >

> 我唯一想到的例外是，一个过程从采样的中间开始(从那时起您就不再有数据了)您将开始直到过程开始.)

> 由于对象太多，我们通常只收集对象\计数器

> > www.kulov.net VSTS MVP, MCT, MCSD, MCPD, INETA Speaker

Is there a bigger goal here than tuning the size of the file?

Every counter gets sampled at the same rate.  The only exception I can think of is that a process that starts in the middle of the sample ( you will have not data from the time you start until the process started).

The best way to control the size of the file is to control the sample interval and object\counter being collected.  Since there are so many Objects we typically only collect object\counter for a specific task.  For example: Looking for classic system bottlenecks I would collect the following every 5 minutes for a week:

o                   LogicalDisk(*)\*

o                   Memory\*

o                   Network Interface(*)\*

o                   Paging File(*)\*

o                   PhysicalDisk(*)\*

o                   Process(*)\*

o                   Processor(*)\*

 

This would generate around 64k to 128k every 5 minutes depending on the number of processes running.

I hope this helps.

Bruce