Performance Metrics - Windows Server

For a VM Server performance check, starts from:

1. Open Task Manager and check for current CPU, RAM utilization. If there is any process you find then check that particular process and dig into it with procmon utility what is doing or depending on the process, follow the further steps to identify what it is doing and why its consuming? The steps may vary depending on the process.

2. If all find then open the resource monitor and find the disk activity. Most of the performance issues causes by the disk IO. Also identify if there is anything.

3. If that is also find then open the command prompt and check the total no of connections on the server by running command:

netstat -ano | find /c /i ":"

You can also check the total no of open connection for a particular port like this:

netstat -ano | find /c /i ":80"

You can also check all the connections on VM with below command:

 netstat -ano > d:\connection.txt

Open this file saved on d: and check the connections and their status. You may find some "Connected". Just focus on "Awaiting Response". They are generally half open connections. If they are not too much then its fine but they are in huge no then check their respective port. If they are in huge no then it means there could be DDOS activity on the port going on which could cause your VM slow down.

These are the few steps which you can follow.

Before digging in to PerfMon I suggest using ResMon.EXE as a starting place.

The Disk tab with the Disk Activity tab sorted via Response Time highest at the top can give a big clue into what's going on. Anything under 25ms to 40 ms is considered reasonable. Anything over 50 ms starts to impact user experience with triple digits being bad.

The other item to look at is the graph on the right. Queue Length, which is Queue Depth in another language ;), indicates the number of I/O operations waiting to be processed by the disk subsystem. Normally, QD can be 2x the number of physical disks in the system to be considered reasonable. Anything over that is an indicator of a problem.

In my experience, the disk subsystem is always the first place to start.

EDIT: If Windows Server Hyper-V is providing the hypervisor then running ResMon.EXE on the host can be helpful while running it in-guest.

Hi Guys,

Thanks for your responses. For what I need to do, I need to have the performance monitored and logged, hence why I'm going to use Performance Monitor. I just need to know what are the best counters are to use for measuring Disk I/O, CPU and RAM.