VMware Cloud Community
goldeneyez
Contributor
Contributor

Super Metric (which created through Custom UI) does not appear on "All Metrics" tab (which exist through vSphere UI)

Hi Everyone

I encounter an issue with new Super Metric creation.

My goal is the following:

I would like to create a super metric which will calculate for each ESXi host server (Separately) the following information:

how much usable CPU capacity and how much usable MEM capacity each ESXi server has.

and I would like to show this information within custom dashboard. and I would like to show this information for each ESXi host server Separately.

I stuck on the initial phase. The new super metric which I recently created does not show up on "All Metrics' tab (under vSphere/Std UI)

I am using VMware vCenter Operations Manager Version 5.8.3 Build 2076729

I have created the following:

1) Environment > Super Metrics > Super Metric Editor

I created new super metric with the following characterized:

First I clicked on the button "This" , so rather than choosing specific Resource , vCops will calculate the Super Metric Separately for each Entity that this Super Metric is attach to.

I choose under Resources any ESXi Sever (not Resource Kinds , because I would like to calculate Separately the Super Metric for each ESXi Host Server).

So the formula finally looks like: $This:M796*0.8

and it's interpreter as This Resource:cpu|toalCapacity_average*0.8

it's 0.8 because I am conserving 20% resources for future growth and for disaster.

my question is: Do I need to choose "Resources" or should I choose "Resource Kinds" ?

The next step I was doing is to create new Super Metric Package.

So I was going to

2) Environment > Super Metrics > Package Editor

I tick the check box which is near the name of the Super Metric which I recently created and that's all. anything else I was not changing. means it's the default configuration.

The next step is to go to

3) Environment > Configuration > Resource Kind Defaults

now within this screen I choose:

Adapter Kind: VMware Adapter

Under Name I choose: Host System

and under Super Metric package I choose: the name of the Super Metric package which I recently created.

The next step is to go to

4) Environment > Environment Overview

and to choose on the left navigation bar: "Resource Kinds". and under it to choose "Host System"

now I highlight all ESXi Host Servers which I would like to apply the recently created Super Metric and click on the edit button.

and for each ESXi host server I attached the recently created Super Metric Package.

in this point as far as I understand , I need to start see under the vSphere UI (means the std UI) , while choosing any ESXi host server from the left navigation pane , and then browsing to:

5) Operations > All Metrics

I supposed to see anything named "Super Metric"

and the problem is: I DONT SEE , the Supper Metric doest exist at all.

what is missing ?

what I am doing wrong ?

thanks in advanced for any assistance.

0 Kudos
9 Replies
mark_j
Virtuoso
Virtuoso

Q: My question is: Do I need to choose "Resources" or should I choose "Resource Kinds" ?:

A: Use ResourceKinds and Attributes for this case.

I'm not sure what the use case is for determining capacity at the Host level in this manner (you're doing the calculation is a very static way), unless of course you don't have clusters (HA/DRS) in your environment. In any event.. use RK and Attributes and you'll be good to go. Make sure to test the SM with the metric visualizer before implementing.. this will safe you much QE time.

If you find this or any other answer useful please mark the answer as correct or helpful.
0 Kudos
goldeneyez
Contributor
Contributor

can you please explain me why I Need to use Resource Kind rather than Resource ?

The Sense telling me that I need to use Resource and not Resource Kind.

even though that when I am using Resource Kind it's seems to work.

0 Kudos
mark_j
Virtuoso
Virtuoso

It's a general rule, and you should only use Resource or Metric when absolutely necessary. Resources and Metrics are absolute references. Whereas ResourceKinds and Attributes are relative references. If you want to refer to a specific resource and metric of that resource, you'd go R#### / M####, which always points back to that SPECIFIC metric of that SPECIFIC resource (statically). Normally you won't do this, and when you do so thinking it's the 'same' as relative reference you'll see some inconsistent behavior of the SM.

Use the relative references wherever possible, the only exception being super metrics and when you create super metrics of super metrics with degrees of separation.

#v5.8.4

If you find this or any other answer useful please mark the answer as correct or helpful.
0 Kudos
goldeneyez
Contributor
Contributor

lets start from zero point.

my goal is to display separately for each ESXi host server the following information:

Physical CPU Sockets Amount

Physical Cores Amount

Total CPU Resources

Physical Memory Amount

Usable CPU Resources

Usable Memory Resources

in last , I will need to generate a custom report which will display the following columns(currently such report design does not exist within vCops):

ESXi Host Name

Physical DataCenter Name

VMware Cluster Name

Physical CPU Sockets

Physical Cores Amount

Total CPU Resources

Physical Memory Amount

Usable CPU Resources

Usable Memory Resources

so , I am right now in the initial phase. first I need to develop the Super Metrics.

so , How to attack the Super Metric configuration ?

my questions are:

1) Should I work with "Resources" (OR) "Resource Kind" ?

2) In case I work with "Resource Kind" , can I use $This rather than $RK## ? again, I would like to display the graph for each ESXi host separately.

3) In case I am using "Resource Kind" , what type of "Looping Function" should I use ?

4) In case I am using "Resource Kind" , when attaching the Super Metric Package , Should I attach it to the cluster level or should I attach it to the ESXi Level ? I would like to get information about each ESXi Server separately on the "All Metrics" tab.

thanks in advanced.

4)

0 Kudos
goldeneyez
Contributor
Contributor

I will be glad if anyone will assist here , thanks.

0 Kudos
mark_j
Virtuoso
Virtuoso

The custom reporting would involve some custom work and some of you're string-based fields wouldn't be feasible in v5.8. I recommend you hold on this report until v6.

The ESXi host metrics you're looking for exist out of the box, no SMs necessary. When you say "usable" on cpu and mem, what exactly are you using for that definition? Overall usable or amount remaining/free? Do you want absolute values or %?

If you find this or any other answer useful please mark the answer as correct or helpful.
0 Kudos
goldeneyez
Contributor
Contributor

Regarding:

Usable CPU Resources

Usable Memory Resources

Regarding the word "Usable"

For each ESXi host Server I would like to find how much resources can be allocated to tenants (VM Machines).

When caluclating the "Usable" Resources I need to work according to the following formula:

Physical Resource which installed on the hardware level

:smileyminus:

How much Resources are allowed to be consumed in maximum

:smileyminus:

[

How much Resources consumed by the ESXi Hypervisor (what also called as "Overhead of the ESXi Hypervisor")

:smileyplus:

How much Resources are reserved for "HA" Cluster

:smileyplus:

How much Resources are reserved for future Growth

]

Now I will explain:

1) Physical Resource which installed on the hardware level - It's self explanatory.

2) How much Resources are allowed to be consumed in maximum - for instance I define that no matter what , the total consumption of resources is not allowed to be over 90% , this means that I save 10% gap.

3) How much Resources consumed by the ESXi Hypervisor - It's self explanatory.

4) How much Resources are reserved for "HA" Cluster - if for example I have a cluster with 2 ESXi server within the Cluster , in such situation I need to save some resources for HA migration , in this example I need to save 50% of the total resources , so in case ESXi Server "A" will be unavailable , so all VM machines will be able to migrate to ESXi Server "B". of course that this is the maximum scenario.

5) How much Resources are reserved for future Growth - for example , I define that on each ESXi Server I would like to save 10% of resources for future growth.

let me know what what do you think about it.

thanks in advanced.

0 Kudos
mark_j
Virtuoso
Virtuoso

If your configuration policies are configured, you could just use "Planning : Capacity Usable "Memory" and "Planning : Capacity Usable : vCPU to pCPU", which is exactly what you want and takes in to consideration the factors you've outlined.

If you find this or any other answer useful please mark the answer as correct or helpful.
0 Kudos
goldeneyez
Contributor
Contributor

Can you provide more information regarding the following built in Metric:

Planning : Capacity Usable : vCPU to pCPU

Because I would like to know how much usable resources I have in terms of (Ghz) or (Mhz). by knowing this data . I can play with the numbers and calculate what ever I want. those numbers will be my baseline for doing any further calculation.

How the suggested built in metric will assist me for doing a capacity planning ? which information this metric reveal ?

thanks in advanced.

0 Kudos