Perfmon进程:处理器时间百分比与任务管理器的CPU使用率(用于监视特定用户) [英] Perfmon Process: % Processor Time vs. Task Manager’s CPU usage (for monitoring a specific user)

查看:362
本文介绍了Perfmon进程:处理器时间百分比与任务管理器的CPU使用率(用于监视特定用户)的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述


我不习惯使用Perfmon和性能监控(所以请放轻松一下吧)


我知道Perfmon没有任何与任务管理器的CPU使用率显示完全相同的东西,但我正在试图弄清楚如何通过Perfmon监控用户的网站CPU使用率(工作进程用户)类似的方式,并尝试理解测量(或如何转换数字以获得类似的理解)


例如,如果在任务管理器中,特定用户一直使用超过5 -10%CPU,我想与用户联系。


我通过示例学习最好,所以这正是我想要做的,具体的例子:



这是在32位双四核Windows 2003 Web服务器(8个CPU)上,服务器上有很多网站,每个网站都有在他们自己的应用程序池/工作进程ID中运行。


通过一些研究我了解到注册表更改(whi) ch我做了)以便PID显示w3wp过程,这样我以后可以通过交叉引用来轻松识别网站。



我使用以下设置设置了一个计数器:

   Process  

- >



处理器

时间

- >
所有实例




这是一个例子。假设我对"黑线"感兴趣下图中的用户流程,因为与其他所有用户相比,他的流程非常高:


http://i35.tinypic.com/106yn8k.jpg




以此为例,我看到此工作进程用户的平均%处理器时间为23.264 ,并且已经飙升至103.124


那么这个23.264数字究竟对我意味着什么?由于此服务器有8个CPU,我应该将此数字除以8吗? (23.264 / 8 = 2.9%AVERAGE CPU LOAD?)


如果这是计算此值的正确方法,这个数字是否与此用户的任务管理器CPU读数的平均值相似? (在任务管理器中,此用户显示为从0到13 CPU负载的任何位置,但在任何高数字时都不一致,似乎2.9%可能是正确的,考虑到周期性峰值的平均值)



我只是不确定我是否正确计算/理解这个计数器,如果这是一个有效的方法来获得关于工人进程用户使用的报告cpu负载过多。



提前致谢。


解决方案

大家好,我注意到很多人都读过这个问题(截至本文的240+),但是如果有人可以发表评论的话,还没有尝试过回答这个问题如果我的问题可能没有正确措辞或混淆?

如果需要,我很乐意澄清我的问题或提供更多细节。

感谢

I'm new to using Perfmon and performance monitoring in general (so go easy on me please ;)

I know that Perfmon doesn't have anything exactly like Task Manager's CPU usage display, but I'm trying to figure out how to monitor a user's web site CPU usage (worker process user) via Perfmon in a similar way, and trying to understand the measurements (or how to convert the numbers to get a similar understanding)

For example, if in Task Manager, a particular user is consistently using more than 5-10% CPU, I would want to contact the user about it.

I learn best by example, so here is exactly what I'm trying to do, with a specific example:

 

This is on a 32-bit Dual Quad Core Windows 2003 web server (8 CPUs), there are many web sites on the server, each running within their own application pool/worker process ID.

Through some research I learned of a registry change (which I made) so that the PID shows up with the w3wp process so I can easily identify the site later by cross-referencing it.

 

I set up a counter with the following settings:

Process
 
->
 
%
 
Processor
 
Time
  
->
 all instances

 

Here is an example. Let's say I'm interested in the "black line " user process in this graph below, as his process is spiking quite high compared to all the other users:

http://i35.tinypic.com/106yn8k.jpg


Using this as an example, I see that this worker process user has an AVERAGE % PROCESSOR TIME of 23.264 , and has spiked as high as 103.124

So - what exactly does this 23.264 number mean to me?  Since this server has 8 CPUs, should I divide this number by 8?    (23.264/8 = 2.9% AVERAGE CPU LOAD?)

If this is the correct way to calculate this, is this number similar to an average of Task Manager's CPU reading for this user?    (in task manager, this user shows as anywhere from 0 to 13 CPU load, but not consistently at any high number, it almost seems that 2.9% is probably about right as an average taking into account the periodic spikes)

 

I'm just not sure if I'm calculating/understanding this counter correctly and if this would be a valid way to get a report on worker process user's using too much cpu load.

 

Thanks in advance.

 

解决方案

Hi guys, I noticed that a lot of people have read this question (240+ as of this post) but there have been no attempts at answering this yet,  if anyone can comment if maybe my question was not properly phrased or confusing?

I'll be happy to clarify my question or provide more details if needed.


Thanks


这篇关于Perfmon进程:处理器时间百分比与任务管理器的CPU使用率(用于监视特定用户)的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!

查看全文
登录 关闭
扫码关注1秒登录
发送“验证码”获取 | 15天全站免登陆