How to diagnose hardware (Storage, CPU, RAM, Network) performance issues Windows Perfmon (4041670)

The scope of this article is meant to provide additional tools that are already available on a windows machine in order to isolate possible performance issues related to:

• Storage
• CPU
• RAM memory
• Network
• Others

Some times it is hard to say just by checking windows event logs or even our AppAssure Support Bundle AAinfo logs, so use this tool wisely and consider scheduling at least 12-24 hours Perfmon Collection Sessions during the peak hours for the Core or Agent Protected Machine in order to be able to pull a relevant sample for the analysis.

For more details regarding Windows Performance Monitor and how to use it to diagnose Server Performance Problems please open these articles:

• Use PerfMon to Diagnose Common Server Performance Problems
• Windows Performance Monitor Disk Counters Explained
• High Avg Disk Queue Length and finding the Cause

Note:  Suppose you have determined the value of Avg Disk Queue Length is indeed accurate and useful – how much is too much? As a general rule for hard disks, an Avg Disk Queue Length greater than 2 (per hard disk) for extended periods of time is considered undesirable. If you have a RAID system with 8 disks, you do not want an Avg Disk Queue Length greater than 16. Faster hard disks with quicker access times (and therefore I/O) will allow greater flexibility with these numbers. Avg Disk sec/read and Avg Disk sec/write should be under 10ms – over 20ms may indicate a bottleneck. If while Avg. Disk Queue Length is over 2 and % Disk Time is hovering at 60% or above, you may want to look into a possible I/O bottleneck.

The Resolution Steps were split into two sections, Collection of Data and Analysis of the Results.

• Collection of Data using the Performance Monitor  - Step by Step Guide:

1. Windows Start Bottom > Run...
2. Type Perfmon.
3. When the Permon Monitor Console shows up, go to Data Collector Sets > User Defined.
4. Right Click on User Defined and then New > Data Collector Set.
   
5. The New Data Collector Wizard stars.
   
6. On the next window chose System Performance and hit Next.
7. Select the Path where the performance data and reports are going to be stored, then hit Next.
   
8.  Use the Option Open properties for this data collector set in order to be able to tune some parameters that are really helpful.
9. Once the Properties Window shows up, open the Stop Condition Tab first, and adjust these parameters, 12 or 24 hours is desired for a good sample during peak hours for the server in question.
   
10. Hit the Apply Buttom.
11. Use the Schedule Tab if it is necessary to start the Collector a a different time, then apply the changes.
    
12. Check the Data Manager Tab and uncheck the maximum folders; edit the Report Name as needed, or use other options as necessary.
13.  Finally hover onto the Data Collector name you just created and hit START.

Analysis of Results

1. Once the collection is completed go to the Performance Monitor Tool using the same steps described on the previous section, then open the Reports Option at the button of the list.
2. User Defined > Then look up the name of your Performance Collector and open it.
3. Review the System Summary Performance first, and validate if CPU, Network, Disk or Memory show any status Busy or something that is not Idle-Normal.
   
4. Validating the Hot Files tab can be very helpful for Repository Issues.
   
5. On Physical Disk Counters, validating Disk Queue Length and Split IO is an excellent indicator of Storage Bottlenecks.
    
6. On Network Interface Counters, validating Packets Discarded or Errors in order to pinpoint possible Network Issues.

      7. You can also pull the Performance Collector Data, compress it and upload it into an FTP site in order to be able to process the data with Performance Analysis of Logs Tool (PAL) later.


The next link Performance Analysis of Logs (PAL) Tool will redirect you to the PAL tool project page where you can find the tool binaries.

The next link Two Minute Drill: Performance Analysis of Logs Tool (PAL) will have instructions on how to use PAL to review the performance logs.