Interpreting vC Ops 5.8.1 vApp Sizing Guidelines

[Mauritz84]

Hi,

We're in the process of provide sizing guidelines to a customer. I'm aware of that it doesn't exist a official sizing calculator. But I need to get some of this stuff clearified.

The official vC Ops sizing guidelines: VMware KB:     vCenter Operations Manager Sizing Guidelines provides you with a good start in understanding the limits. But this guidelines get me confused. It tells us that the VMs supported is 12 000 for the full profile. But in the 5.8.1 release notes it's stated that 12 000 objects is supported. Objects is as you know VMs, Hosts, Datastores etc. All of these have metrics. So if we would have 12 000 VMs there would also be between 300-500 hosts depending on your ratio and you could expect the double amount of datastores.

Doing the calculation with average metrics values with the above object numbers it's getting really near or above the 5M metric limit and the 12K object limit of the full profile for the vApp.

This depends ofcourse which average value you'll use.

12000 * 300 (VMs) + 500 * 1500 (Hosts) + 1000 * 250 (Datastores) = 4 600 000

Does anyone know how this is supposed to be interpreted?

Regards,

Erik Alm

[mark_j]

The emphasis is on VMs representing objects because they represent the largest part of the resource population as well as metrics (typically).

12,000 is the recommended max because of the overhead that is placed on the analytics and ui components.

Interpret it as 12000 generally supported max # resources, regardless of profile you're using (virt hardware reqs will vary, obviously). If you need to go above that #, give GSS a call and explain the situation and they'll do some advanced sizing calcs to get you to where you need to be as well as known caveats. However, at 12000+ resources you'd likely want to begin considering splitting off different vC Ops instances based on vCenter in order to scale out better with vC Ops 5.x.