In my 1-Click Capacity Planning vCOps Dashboard blog series I’ve created 6 vCOps Super-Metrics. Two of them calculating ESXi host cluster CPU and memory total provisioned capacity. But let us not forget HA Admission Control behind.
Looking at HA Admission Control policy numbers in a capacity planning process is more than a valid point. In the world of vCOps, configuring the policy to take HA under consideration is super important and will have effect on vCOps capacity planning module results. I had a blog post a few months ago around vCOps usable capacity.
So, what about creating some Super-Metric which will respect Admission Control reservations?! In order to do so, I will be using vCOps 5.8 and will assume that Admission Control policy set to reserve 10% for CPU and 25% for memory.
Total CPU Usable Capacity in GHz (With HA Admission Control at 10%) |
---|
($This:cpu|capacity_provisioned/1000)-(($This:cpu|capacity_provisioned/1000)*0.1) |
cpu|capacity_provisioned = Resources > Cluster Compute Resource > CPU Usage > Provisioned Capacity (MHz) |
Total Memory Usable Capacity in GB (With HA Admission Control at 25%) |
---|
($This:mem|host_provisioned/1048576)-(($This:mem|host_provisioned/1048576)*0.25) |
mem|host_provisioned = Resources > Cluster Compute Resource > Memory > Provisioned Memory (KB) |
Now that we have our cluster usable capacity Super-Metrics set we can use it in our capacity planning dashboards.
But wait, what about showing the actual reserve % of CPU and Memory within a Scoreboard widget? Lets create couple more Super-Metrics:
HA Admission Control Total CPU Reserved = 10%) |
---|
((0.1*($This:cpu|capacity_provisioned))/($This:cpu|capacity_provisioned))*100 |
cpu|capacity_provisioned = Resources > Cluster Compute Resource > CPU > Provisioned Capacity (MHz) |
HA Admission Control Total Memory Reserved = 25%) |
---|
((0.25*($This:mem|host_provisioned))/($This:mem|host_provisioned))*100 |
mem|host_provisioned =Resources > Cluster Compute Resource > Memory > Provisioned Memory (KB) |
Now that we have “Admission Control Super-Metrics” set, we can use it in our dashboards. My favorite for this one is the Scoreboard widget; you can use either the interactive way using XML’s or just create a static widget like this by selecting a specific cluster.
I’m enjoying your vCOPS capacity planning series of blog posts
Is there a way to automatically calculate HA % based, so as hosts are added the supermetric automatically updates to stay allowing for N-1 ?
Thanks Troy!
Currently there is no way automatically calculate HA admission control in vC Ops.
Thanks for all your work, it helps out tremendously. I have a question. Before I even enter the percentages for HA my calculation for total capacity isn’t correct from what I can tell.
cpu|totalCapacity_average = Resources: Cluster Compute Resource > CPU Usage > Total Capacity (MHz)
When I look at a given cluster say ClusterA, my metric graph shows 132.37 Total CPU Capacity, when I look in Vmware I get 153 Ghz on the summary page for the cluster. I double checked the formula and it looks exactly like your example. Any reason for the difference between the metric and what Vmware shows on the cluster summary page?
Thanks,
Greg
Thanks Greg!
What version of vC Ops you are using?
Notice I have cpu|capacity_provisioned in my formula and not cpu|totalCapacity_average.
5.8.1, I’ll take a look again.
In your other three part series you use capacity avg vs provisioned capacity. Why?
Cluster Total CPU Capacity (GHz)
sum(This Resource: cpu|totalCapacity_average)/1000
cpu|totalCapacity_average = Resources: Cluster Compute Resource > CPU Usage > Total Capacity (MHz)
When I select Resource Kinds, Cluster Compute Resource and My Cluster under Super Metrics I am not seeing the two metrics I created in your example in my widget? Any help would be appreciated?
Thanks,
Greg