Alert for Virtual windows server CPU

Hi,

if I understand correctly, you are interested to CPU utilization for whole virtual machine, not for individual cores.

Please follow these steps or import attached alerts in ZIP file 'Alerts_CpuUtilization.zip'.

Virtual machine is managed as a node - data are polled e.g. through WMI

1. Create new alert
2. Tab trigger condition
   • Type of Property to Monitor: Node
   • All conditions have to be met, so select 'Trigger Alert when all of the following apply'
   • Add these complex conditions
     • filed Node name is equal to value <name of virtual machine>
     • field Node ID is equal to value <node id of virtual machine>
     • field CPU Load is greater than value <e.g. 85>
3. Tab reset condition
   • Let checked 'Reset when trigger conditions are no longer true' or create your own reset condition. Similar to trigger condition.

• You can not use data from table 'CPUMultiLoad'. There are data for each virtual machine's core separately.
• If you are interested for all virtual machines, let out filtering conditions - Node name and Node ID.
  

Virtual machine is polled through vCenter - data are polled through VMware API

• There is no need to have virtual machine managed as a node, but parent ESX host has to be managed as a node.

1. Create new alert
2. Tab trigger condition
   • Type of Property to Monitor: Virtual Machine
   • All conditions have to be met, so select 'Trigger Alert when all of the following apply'
   • Add these complex conditions
     • filed Display name is equal to value <name of virtual machine>
     • field Virtual Machine ID is equal to value <id of virtual machine>
     • field CPU Utilization is greater than value <e.g. 85>
3. Tab reset condition
   • Let checked 'Reset when trigger conditions are no longer true' or create your own reset condition. Similar to trigger condition.

• If you are interested for all virtual machines, let out filtering conditions - Display name and Virtual Machine ID.

Best regards,

Michal

Hi,

Thank you so much for your reply and I really appreciate your help.

I have tried both queries and in Alert Manager but not getting the alerts. I have tested those queries in alert manager and the results are as follows:

Test results for query 1

{

Virtual machine is managed as a node - data are polled e.g. through WMI

1. Create new alert
2. Tab trigger condition
   • Type of Property to Monitor: Node
   • All conditions have to be met, so select 'Trigger Alert when all of the following apply'
   • Add these complex conditions
     • filed Node name is equal to value <name of virtual machine>
     • field Node ID is equal to value <node id of virtual machine>
     • field CPU Load is greater than value <e.g. 85>

}

Alert: CPU on the <node name> is = 56 %

Alert Name: Test alert for <node name> CPU utilization

So it's not giving me the the right answer because cpu 1 =56% and cpu 2 = 50% - I am looking for alert email for combined utilization which would be 106%

Test results for query 2

Virtual machine is polled through vCenter - data are polled through VMware API

• There is no need to have virtual machine managed as a node, but parent ESX host has to be managed as a node.

1. Create new alert
2. Tab trigger condition
   • Type of Property to Monitor: Virtual Machine
   • All conditions have to be met, so select 'Trigger Alert when all of the following apply'
   • 
     
   • Add these complex conditions
     • filed Display name is equal to value <name of virtual machine>
     • field Virtual Machine ID is equal to value <id of virtual machine>
     • field CPU Utilization is greater than value <e.g. 85>

Alert: CPU on the <node name> is = ${CPULoad}   (I am not sure why its not returning any number here - it might be because virtual machine managed as node - not sure)

Alert Name: Test alert for <node name> CPU utilization

Thanks & Kind Regards

KP.

Hi,

now I think that understand your requirements.

I have updated alert for nodes, please verify.

Best regards,

Michal

Hi,

it is attached as *.AlertDefinition file in my previous comment.

I have zipped that and attaching it again as 'NodeAlert_CpuCoreUtilization.zip'.

Best regards,

Michal

Hi,

I got the files - for some reason it's not showing in the inbox of my account - (really weird) but I went latest post on website - I see the files.

Let me check and get back to you. Thanks.

Hi,

This query is working. Many Many thanks to you!

I have one more request - Can you define SQL variable for alerts, so that we get the value of CPU 0 and CPU 1 or sum of it.

I am not SQL expert but I can modified query as per my needs.

Please see the below image and if you can define SQL variable for email.

I really appreciate your time and help!

Thanks & Kind Regards,

KP.

Hi,

I updated alert and added variables for logging. There is created action for logging to NetPerfMon EventLog. Similary you can create email action.

Variables are:

• CPU Cores Utilization - sum of all CPU cores utilization,
• CPU Cores Count - number of CPU cores.

I attached alert in zip file 'NodeAlert_CpuCoreUtilization_v0.2.zip' or just insert these commands:

Node CPU Cores utilization

Node ID: ${NodeID}

Node Name: ${Caption}

CPU Cores Utilization: ${SQL:SELECT SUM(CPUMultiLoad_Detail.AvgLoad) AS CPUCoresUtilization FROM CPUMultiLoad_Detail WHERE (DATEDIFF(mi, CPUMultiLoad_Detail.TimeStampUTC, GETUTCDATE()) < 5) AND CPUMultiLoad_Detail.NodeID = ${NodeID}}

CPU Cores Count: ${SQL:SELECT COUNT(CPUIndex) AS CPUCoresCount FROM CPUMultiLoad_Detail WHERE (DATEDIFF(mi, CPUMultiLoad_Detail.TimeStampUTC, GETUTCDATE()) < 5) AND CPUMultiLoad_Detail.NodeID = ${NodeID}}

Best regards,

Michal