When it comes to troubleshooting in a virtualized environment, the component which is shared most often between the hosts, is most likely a cause of a performance problem. If you look at raw performance metrics you may ask yourself, is it good or bad, normal or abnormal. Let me give you an analogy. If you take a picture of something and the size of the object you’re shooting isn’t well reflected and you need a context, you can put a coin aside the object and take another picture. Now you got a context and you can better imagine the size of the object on the photo. Well, vRealize Operations is giving you exactly this kind of context for all your Datacenter / IT metrics you have to worry about. This post is about how to enable Datastore and Storage related Workload metrics like IO per second, throughput and latency for VM and Datastore objects in vRealize Operations 6.1.
Lets take a look at the first two pictures. In the 1st image you can see that the Workload tab is grayed out for the Datastore. The 2nd picture reveals the same. No Datastore or IO metrics. Only CPU and Memory is displayed for the VM object.
To make IO metrics available we just need to adjust the vROps policy. Or better, create a new policy that inherits all the base settings from the default settings plus enabled storage Workload metrics for VM and Datastore objects. To do this, click on Administration -> Policies -> Policy Library -> Add.
We don’t have to edit all the sections for our new policy, therefore I walk you only through the necessary steps to create the policy we need. First section, what policy should be the starting point. In other words what will be the parent policy. I selected the vSphere Default policy, which is created when joining the vCenter to vROps. Jumping to the 3rd section. Now we choose what object types we want to change. As I said, I want to include IO Workload metrics for VM and Datastore objects. So I choose vCenter Adapter – Datastore and vCenter Adapter – Virtual Machine.
Now we can can select the relevant metrics that should be visible for Datastore and VM objects. Open the Workload section and tick the boxes for the metrics you want to include. I’ve chosen all IO relevant metrics as you can see in the screenshot. I also enabled those metrics for the overall stress calculation of the given object. Repeat this steps for the vCenter Adapter – Virtual Machine section.
Nearly finished, jumping to section 8. Here we have to choose where the new policy should be attached. In my case, i just want to replace the default policy with the newly created one. Because we selected, that the new policy has to inherit everything from the vSphere Default Policy we are good to go. I selected vSphere World and saved the policy. Depending on your configuration and the complexity you have to face this step could be other than described here. Last but not least, I’ve chosen that the new policy should be the new default policy for all objects in my environment.
As you can see in the screenshots, the relevant storage metrics are now displayed for the VM Object and the Workload tab for the Datastore object can be selected for better troubleshooting.
Final conclusion. Making storage relevant metrics available definitely helps to troubleshoot storage related issues in your virtualized environment.