Dashboard 4.19.2. New Used/Allocated resources view ???

[SviridoffA]

Hi everyone! Is anyone else experiencing issues with the new resource display, with the new Used/Allocated toggle? I’m completely lost as to what it’s showing. I have several virtual machines running on my cluster. They’re almost constantly idle. However, the dashboard, when set to “used,” displays something similar to my “allocated” values. Meanwhile, “allocated” shows values that don’t make sense to me. My current workload on the cluster should be around 1%, and “allocated” should be closer to what is currently displayed as “used.” I t is not fresh install, I upgraded from 4.19.1. I’d appreciate any help in understanding this.

[nvazquez]

Hi @SviridoffA thanks for your question. The toggle was introduced to fix a mismatch on the previous versions about the used vs allocated values that were displayed on the dashboard, allowing CloudStack to display allocated and used capacity separately. Previously, ony allocated values were meant to be displayed, but CloudStack was using the used capacity values. This was fixed on this PR: #10215. According to your images it looks like there is a problem with the CPU used capacity values displayed on the dashboard (according to your cluster stats it should be 2% instead of 22%). I'll have a look at this

[SviridoffA]

Hi @nvazquez thanks for your reply. I think this is a great feature, but unfortunately it doesn’t seem to be working in my case. Please let me know if there’s anything else I can provide. I’m not sure if it’s relevant, but I do have overprovisioning configured as written below:

CPU - 2
MEM - 3

And one more thing.. It looks like the allocated amount is showing up as a few percent higher than it really is. Currently, my workload is very static; sudden spikes are almost non-existent.

[DaanHoogland]

@SviridoffA , I do not recall, did you create an issue for this?

[SviridoffA]

No, I didn’t create one because I wasn’t sure if it was just my problem, but let me do that now.

[SviridoffA]

addressing this: #10803
and closing this one

[DaanHoogland]

neat

[weizhouapache]

@SviridoffA
The values are fetched via cloudstack API, it is not a UI issue.

Moreover, I think it is not a bug. for clarification

• On the dashboard boarding page, data are fetched by API listCapacity

  • Allocated capacity: the resource allocated to the user vms and system vms, etc
  • Used capacity: the resource used by Running vms, or reserved for vms which are stopped recently

• On the Cluster/Zone details page, data are fetched by API listZonesMetrics or listClustersMetrics

  • Used: the real usage of hosts in the zone or cluster. It is consistent with the usage users get on the host (for example, free, htop, uptime)
  • Allocated: the resource used by Running vms, or reserved for vms which are stopped recently (same to Used capacity above) in precentage
  • if the used percentage is much smalled than Allocated percentage, it means you could allocate more resources on the environment. You can increase the over provisioning factor (be careful on MEMORY if you use KVM).

@nvazquez @DaanHoogland

[SviridoffA]

Before deploying and running a VMs as a cloud provider, there is a process ofresource planning. The thing is, in today's world, without overprovisioning, a service provider often becomes unprofitable. To make resource planning more convenient, it would be great if the root admin dashboard displayed statistics on the current load on the cluster(s) and the total amount of allocated resources, with total available. This would allow for a rough forecast of what overprovision ratio could be used. Honestly, when I first saw this toggle, I thought that’s what it was for. Obviously, I’m missing something. But, current display doesn’t make any sense to me. Perhaps the toggle should have been labeled as reserved/allocated instead of used. But I still don’t understand what this metric is for. Don’t get me wrong, I’m not saying it’s incorrect. I’ve been working with CloudStack for a little over a year, and I’m probably missing something. Maybe I should dig into this call through the API manually, and I’ll figure out what it returns. But right now, the CPU and memory numbers shown in the dashboard don’t align with any real metrics for me. Maybe I just can’t correlate them properly.

[SviridoffA]

@weizhouapache Can you confirm please, that these are the correct metrics. Just to be sure)))

[SviridoffA]

It seems I finally got it . I initially misinterpreted the switch's position, and it was also confusing that the metric also accounts for recently stopped virtual machines and doesn't update immediately.
I have tested it by creating a large number of VMs, and now everything makes sense to me. I apologize for the mess I caused here.
@weizhouapache Thanks for you clarification. Only one question for now, Can you tel me please, how long resources are reserved for stopped vms?

[weizhouapache]

It seems I finally got it . I initially misinterpreted the switch's position, and it was also confusing that the metric also accounts for recently stopped virtual machines and doesn't update immediately. I have tested it by creating a large number of VMs, and now everything makes sense to me. I apologize for the mess I caused here. @weizhouapache Thanks for you clarification. Only one question for now, Can you tel me please, how long resources are reserved for stopped vms?

no problem @SviridoffA
it is great everything is clear to you. hope other users can get some more information from our discussion

how long resources are reserved for stopped vms?

it is configurable. There is a global setting capacity.skipcounting.hours
Capacity skipcounting hours (capacity.skipcounting.hours): Time (in seconds) to wait before release VM's cpu and memory when VM in stopped state
The default value is 3600 seconds (1 hour)

[SviridoffA]

Thank you once again, closing this one...