Glad to see that vRealize Operations 6.5 sports the ability to import & export Groups. This makes it much easier for you to replicate the Operationalize Your World dashboards into your environment. The entire chunk of group creation is no longer required!
Starting from 6.5, you can also import & export multiple groups. No need to create it one by one. The import dialog box is similar to other import dialog boxes.
The group description is actually pretty simple. Let’s use an example to show it. This is what you see on the UI. It’s a group called Idle VMs that grabs all VMs whose CPU Idle Time is greater than 99%.
Here is the JSON file. I’ve highlighted some of them, so you can see how it’s mapped.
I’ve been using 6.5 release and generally found it stable. There are also a number of bug fixes addressed in this release.
As not all customers are on vRealize Operations 6.5 yet, I’m keeping the instructions to create the groups in this article. Use it as the guide if you are running 6.3 and 6.4. If you are on 6.5 or higher, refer to the simplified step here.
I have also simplified the steps by manually creating a dummy policy to help you do bulk import. As a result, the video guide is no longer required. I’m keeping the videos here just in case you need it.
Create the following groups. All of them. Note it’s case sensitive! If you do not use these names exactly, your dashboard will get hourglass icon.
- Under the group type Class of Service
- Tier 1 (Gold)
- Tier 2 (Silver)
- Tier 3 (Bronze).
- Under the group type Function
- Datastores (Shared)
- Datastores (Local)
- Under the group type VM Types
- Idle VMs
- Large VMs (CPU)
- Large VMs (RAM) –> do note this requires vSphere 6U1
- Powered On VMs
- Powered Off VMs
- VM with no VMware Tools
- VM with VMware Tools installed
- Under the group type Tenants
- Tenant ABC, etc.
- [e1: this part is optional. Create 1 group per tenant. See this for details]
Once created. Do not rename the object.
For each tier, ensure you select the right Cluster, VM and Datastore that you have planned earlier. Do not do an impromptu planning. That’s an oxymoron 🙂
You need to select these 3 objects. If you do not select the object, you cannot apply the Performance SLA.
For datastore, exclude local datastore unless they are part of your official Service Tier. vR Ops 6.3 has a property for that. Pretty cool!
In any group, always do a preview before you save. Take note the total number of member.
For the Idle VMs group, define what suits your operations. I use >99% CPU Idle Time. This is based on a 30-day period (default setting in the policy), so that translates into maximum of 7.2 hours in the last 30 days.
The default setting in policy is 90%, which differs to what I use in the group. Does it mean you do not have to worry about the Policy Settings for Idle?
You are right! Don’t worry about the policy. I do not use Is Idle metric. I use the CPU Idle Time metric. No need to modify the default settings as you’re not using it.
For the Large VMs groups, you can change the definition to suit your need. I’d recommend changing below from 4 vCPU to at least 7. If you have a lot of 8 vCPU VMs, then change to 8 so they are not included. Focus on the big ones.
Make sure you only choose powered on VMs too, else they get added! See below on how to add this condition. You can use this metric, or use the Summary | Running metric. For RAM, I use in-guest metric as it’s more accurate. Just because a VM is powered on, does not mean the OS is running.
For the Powered Off VMs group, I define them as VMs that are off for >50% in the past 30 days and they are powered off at the moment. This is conservative, as that VM needs to be powered off for a total of 15 days in the past 30 days.
For group of VM with VMware Tools installed and group of VM with no VMware Tools, use the property shown below. I cloned the group, and simply change from is to is not.