Kaseya Community

Windows Performance Counters in Network Monitoring Vs Core Monitoring

This question is not answered

So I have been tasked to create Hyper-V monitor sets as the ones we currently have in the VSA core Monitoring is lacking Network monitoring because core Monitoring for some unknown reason is incapable of Monitoring Network Interfaces.

To reconfirm this I created a core monitor set for Hyper-V network performance counters;

And as expected I got absolutely no data from any of these counters (@Kaseya still waiting an explanation on why this does not work...).

Kaseya's Network Monitor is suppose to overcome this issue so yesterday I started working on this in Network Monitor (R9.1) and noticed something strange.

When I try and add a Windows Performance counter I only get 2 Objects to select from. I checked this against several Hyper-V hosts and got the exact same 2 only Performance counter objects so the missing counters that I'm expecting to see is no just limited to one Hyper-V Host.

I tried to rescan however it did not add any more objects and if you compare the list of Hyper-V Performance Counter object in Kasey core Monitoring, most that I can currently use something is definitely wrong here.

Also considering that this is a Windows machines I'm expecting to see many more Performance Counter Object not related to Hyper-V yet they are not there.

I have looked at all the resources I could find online, but it I have so far found nothing that can help me figure this out, I have even re-watched the "Kaseya Support TechJams: KNMi Overview & Troubleshooting" but unfortunately they don't cover Windows Performance counters in it at all;

Kaseya Support TechJams: KNMi Overview & Troubleshooting
https://www.youtube.com/watch?v=zXNBOkNw_DE&list=PLyTkRXb2tZwYSCkLaGWqonPiBufhQ3UKX&index=5

All Replies
  • Hi  

    Have you created a support ticket on these issues?

    If you have not, I would strongly recommend creating a ticket for both the KNM issue and Classic Monitor issue you are having.

    Have you tried to monitor this Hyper-V environment using a CIM monitor opposed to Performance Counter within KNM?

    I am not sure if CIM will have the same monitor options you are looking for from Performance Counters, but it may be worth a shot.

    Let me know either way.

  • Just an update on what I have learned so far, I have found that our VEEAM backup server that has the Hyper-V role installed displays all of the expected performance counters.

    So I created the monitor sets on this server, templated them and successfully applied the majority of them to the Hyper-V host where only the 2 counter objects were visible.

    The counter objects that did not work, was because the instances counters did not exist or so I thought. An example of this was the Hyper-V Virtual Network Adapter (Similar to the Network Adapter Performance Counter), it has required counter instances that you have to pick from and that appears to be specific to the computer you create the monitor set on (or same hardware);

    After clearing the counters from the VEEAM backup server and reapplying the same template I created using this server with counters that had a counter instances specific to this server Alarms with reason that the "Instance ... not found". This should have worked as I used this server to create the template in the first place.



    Note that If I manually recreate this performance counter with the same instance name it works without any problems.


  • Hi Nicolas Ponce

    I have created a ticket a very long time ago for the Classic Monitor issue where Network Interface counters do not generate any data and they told me that it is a known issue and that they would fix it in a future release. I suspect KNM was their fix however I still see no documentation or any written material from Kaseya why it is not working, I would have at least expected a small note about it in the help document to warn customers about the issue.

    Also, no I have not created a ticket for the KNM issue because;

    1) Historically I found that your Tier 1 support was unable to help me with similar issues in the past and when they escalated the tickets I often got either bad/wrong information, they require access to our VSA portal which is not possible at this point in time or they took week/months to respond by which time have figured out a solution or a workaround myself.

    2) I'm unsure if the issue is known or not often the fastest way to find out, is to post to the community forums and see if anybody else has run into the same issue.

    I have not tried CIM monitoring yet, to be honest I have no experience with monitoring via CIM as I have been using Kaseya for the last 8 + years and KNM has only became available to me this year in our test environment.

    The reason I'm using Performance over CIM monitoring is because all the available monitoring information out there provided by Microsoft and other Monitoring software deals with Eventlog, Service and Performance based monitoring.

    Why do you want me to use CIM monitoring over Windows Performance monitoring?

    I tried the CIM monitor against several server I have been applying the performance counters too and found that they do not show any "Class" options even if I try to rescan;

    I suspect the reason this is because I need to open ports on the firewall to allow this.



    typo
    [edited by: HardKnoX at 7:21 PM (GMT -7) on Oct 15, 2015]
  • The "LogicalDisk::% Disk Time % Disk Time" Performance counters does not seem to work in either KNM and core Monitoring.

    Has anybody else had any success with "LogicalDisk::% Disk Time % Disk Time" Performance counters?

  •  

    Apologies for the delayed reply here, it has been some time since I could sit down and get back to many threads.

    To address or possibly assist with some of your issues in line - (I did not include screenshots to save space)

    I have created a ticket a very long time ago for the Classic Monitor issue where Network Interface counters do not generate any data and they told me that it is a known issue and that they would fix it in a future release. I suspect KNM was their fix however I still see no documentation or any written material from Kaseya why it is not working, I would have at least expected a small note about it in the help document to warn customers about the issue.

    Also, no I have not created a ticket for the KNM issue because;

    1) Historically I found that your Tier 1 support was unable to help me with similar issues in the past and when they escalated the tickets I often got either bad/wrong information, they require access to our VSA portal which is not possible at this point in time or they took week/months to respond by which time have figured out a solution or a workaround myself.

    2) I'm unsure if the issue is known or not often the fastest way to find out, is to post to the community forums and see if anybody else has run into the same issue.

    Understood, this must have been prior to Kaseya moving from Service Desk to ZenDesk ticketing system (Roughly 2~ years ago).

    Most likely, the "I Suspect KNM was their fix" comment referred to could indeed be a fix - but not necessarily a fix using "Perfmon Counters". If you still have the ticket # on this case I can dig it up in the old Service Desk platform (portal.kaseya.net).

    I understand the frustrations around dealing with support as I used to work on the Support Team for several years and now work with them to ensure issues are resolved.

    Also, I definitely agree that the Kaseya Community provides invaluable knowledge and one of the fastest ways to acquire this information.

    With that said, If you do decide to submit a ticket and have issues with support - I strongly encourage you to reach out to me or any of the other Kaseya resource on the community (Feel free to reach out directly via PM, Email, or a Thread)

    We can ensure that you get the proper support you require.

    I have not tried CIM monitoring yet, to be honest I have no experience with monitoring via CIM as I have been using Kaseya for the last 8 + years and KNM has only became available to me this year in our test environment.

    The reason I'm using Performance over CIM monitoring is because all the available monitoring information out there provided by Microsoft and other Monitoring software deals with Eventlog, Service and Performance based monitoring.

    Why do you want me to use CIM monitoring over Windows Performance monitoring?

    I tried the CIM monitor against several server I have been applying the performance counters too and found that they do not show any "Class" options even if I try to rescan;

    My intention was not to force you to use CIM but inquire if you have attempted to use this function opposed to Performance Counters.

    CIM was built as the standard which is what WMI / MI is built on top of and provides several monitoring options for unique environments (Especially Hyper-V and ESXi hosts).

    https://msdn.microsoft.com/en-us/library/aa384642(v=vs.85).aspx

    https://msdn.microsoft.com/en-us/library/aa389234(v=vs.85).aspx

    However, I still believe the issue you are reporting needs to be identified and clarified by our Support/Engineering team (Which I am more than happy to assist with).

    I suspect the reason this is because I need to open ports on the firewall to allow this.

    Perhaps, you will also require CIM Credentials enabled in KNM to properly poll a device.

    You can test connectivity, user credentials, and some of the available classes using WBEMTest.exe for several WMI/CIM functions:


    This program should be available on all windows OS types.

    The "LogicalDisk::% Disk Time % Disk Time" Performance counters does not seem to work in either KNM and core Monitoring.

    Has anybody else had any success with "LogicalDisk::% Disk Time % Disk Time" Performance counters?

    Are you referring to Performance Counters on the Hyper-V agent itself or in general, all %Disk Time counters for all machines?

    If you are not getting data from the Core Monitoring > Logical Disk>  % Disk Time counter, check to see in <installdrive>\kworking folder\klogs folder if the .CSV file is returning data:


    I know this is all very late, and you may have resolved this at this point and I hope it provides some sort of helpful information.

    I will be standing up a Hyper-V host and checking out the KNM/Core Monitorings Performance Monitor counters against it to review the behavior as well.

    When I am at the point to create a ticket, if you have not, I will do so and update this thread with the ticket number and information I get as the investigation moves along.

  • Hi Nicolas;

    Thanks for your detailed response, I will go through it and update you as soon as I can.

    I have not yet resolved it as I'm making a list of all the counters that I can't get to work and then will check how important these are for what we need them for and I will log the tickets as needed.

    Here is something that might be of interest, the DLINK Memory utilization OID in the supplied KNM DLINK switch template appears to be invalid;

    I suspect that the OID that is displayed is actually the Memory utilization value as in that the example above's Memory utilization is 97%.

    I did some research and think the OID for Memory utilization for the DLINK switches should be;

    1.3.6.1.4.1.171.12.1.1.9.1.4.1

    When I replaced it, the Memory utilization object works,

    I unfortunately have no way to update the templates that are provided with KNM so this is something you guys will have to fix. My plan is to create a separate DLINK monitor template with the updated OID and a few additional ones that is not part of the supplied template.

  • Apologies I have not been able to make any headway on the %Disk Time% counters yet.

    My employer recently went to one of the Kaseya social events and was sold on the idea to become part of the Kaseya Beta program which forced me to upgrade our test server from R9.1 to R9.2 beta.

    After upgrading most of the devices I had configured in KNM for monitoring now has SNMP and performance counters that no longer work.  (Ticket #108431)

  • Hey  

    Not a problem at all. I have been trying to stand up my own environment to mimic this issue in the interim.

    However, I can possibly help on your specific issue in the ticket referenced.

    I am going to email you directly.

    Best,

    Nicolas

  • Hey HardKnoX, Nicolas,

    I've read above....." is to create a separate DLINK monitor template with the updated OID and a few additional ones that is not part of the supplied template. "

    would you be able to explain as to how create temporay own KNM Templates

    rgds

  • Hi RDK;

    Your question is a bit difficult to understand, it sounds like you are asking how to make Monitor templates in KNM so I will go with that.

    First you need to find an existing device that you want to monitor, you create new monitor for the device using the "Add new Monitor" from the "Commands" menu.

    Once you have created all the monitors you need select the device and look at the "Commands" menu option called "Save as template"


    After you have save the template it will then be select-able from the "Apply templates" option from the Command Menu and you can also access it from the "Settings > Device Templates", your custom templates will be Hyperlinked

    There are a few KNM videos on Kaseya's Youtube channel that is also accessible under the Kaseya Community Resources that you might want to look at if you haven't already.

  • Hey  

    Finally circling back to this topic, sorry for the delay.

    I stood up a Hyper-V instance (No GUI) on a Dell box a few weeks ago and have been spending some time understanding some of the differences.

    So, to clarify what OS / Version I am using:

    Immediately ran Update Lists by Scan to populate the detected Performance Counters.

    It brought back a list very similar to what you are seeing.

    I added various performance counters from the Classic Monitor module to test what would populate (some were exactly like the ones in your image, or the best I could tell from the image).

    Some reported data and responded back and some did not:

    So, I took a look inside the Hyper-V machine to see what was actually returning data and what actually existed on the machine.

    You can see below that the "Legacy Network Adapter" Counter is not returning any information and it appears the counter does not exist on this Hyper-V instance while the Hyper-V HyperVisor Logical Processor counter is collecting data and appears to exist properly:

    Going further into the problem, I ran a typeperf -qx and exported the results to a text file.

    I have made the results available at the following link:

    https://paper.dropbox.com/doc/typeperf-qx-CEh9C6GsP4EcebUc6ovhs

    So, from what I can see Update Lists by Scan is detecting these counters as part of the OS image and populating them in the VSA.

    However, either my instance or all instances require some action or pre-requisite for this monitor to actually populate data (Example: Legacy Network Adapters).

    Unfortunately, my instance is very bare bones - so perhaps some of this information may assist you in reviewing the behavior for your classic monitor performance counters.

    In regards to what KNM is detecting, I get the same results/behavior you do when trying to use Performance Counter monitoring on this box within KNM:

    This may or may not be related to the architecture of the device. 

  • Hi Nicolas;

    Thanks for the info.

    In a nutshell I think we have identified two potential KNM software bugs.

    The majority of the Hyper-V hosts that I have tried to use agentless Performance monitoring against are actually Standard or Datacenter edition Windows  2012 R2 servers same as the rest of server builds we use. These servers are configured with either minshell or full GUI mode so having the server in core mode I suspect is not the cause of the issue of the missing performance objects.

    The only server that I could see all of the counter object on was a VEEAM backup server (Windows Server Standard Edition with full GUI mode) which also happen to have the Hyper-V role installed. When I have some more time I will try and figure out why this VEEAM server works and the other Hyper-V hosts don't.

       

    The Hyper-V Network Legacy Adapter I think need to be in use on one of the VMs before the counters can generate data and its commonly used on non-Windows based VM's.

    The problem here was specifically with the Hyper-V Virtual Network adapter counter object did not work after the monitor set was template'd and reapplied to the same or other Hyper-v servers.

    This appears to be a common problem with performance objects that have uncommon counter instances, in this example it was the Realtek Adapter device. The secondary part of this issue for this example is that you can't modify the instance and the only way that I could get it to work was to create a fresh performance object with the exact same settings defeating purpose of templating it.

  • Sorry to resurrect this post.

    I'm in the midst of testing out HyperV monitoring in an environment and have come up against the same issues. Has any headway been made on this issue?

  •  

    What are your specific issues?

    Are you trying to use a template for Hyper-V performance objects?

  • Hi Nicolas,

    I can't seem to see the performance counters on a server with the Hyper V role installed.

    I only see Hyper V Virtual Machine Health Summary