VMware’s vRealize Operations Manager is a powerful analytics platform that helps administrators gain insight into their application, virtual, and physical infrastructure. One challenge you might have as you deploy vRealize Operations is the sheer amount of data that you throw at it, and the CPU, memory, and disk requirements to handle the collection, analyzing, and storage of that data. In this blog, I’ll explore a few options for “dialing in” the metrics that you collect with vRealize Operations – to tune out the metrics you could care less about in order to gain greater capacity and performance within your vRealize Operations cluster.
There are a few options we will explore for limiting resources and metrics are collected. The first two will vary depending on the vRealize Operations Management Pack that you have deployed. The third method uses vRealize Operations Policies to restrict what metrics are collected and how they are processed.
The first option is to leverage the Advanced Configuration for the Adapter Instance configuration. The options available here will vary depending on the Management Pack. For the VMware vSphere Management Pack (figure 1), you can disable vSphere Distributed Switch, Virtual Machine Folder, vSphere Distributed Port Group from collection. Additionally, you can exclude Virtual Machines from Capacity Calculations, resulting in less analytics overhead. Simply toggle these values and then press Save.
Figure 1: Adapter Configuration – VMware vSphere
Other Management Packs will have different options. For example, the Cisco UCS Management Pack from Blue Medora includes the ability to monitor down to the physical and virtual port (vHBA/vNIC) level. You can optionally turn off port monitoring in order to save resources. (Figure 2)
Figure 2: Adapter Configuration – Cisco UCS
A second option available with some Management Packs is the ability to set similar configuration in a text file. To make changes to these text files, you will either need to open up the vROps appliance in the vSphere console or enable SSH on the vROps appliance and SSH in as the root user. For this example, we’ll take a look at the NetApp Storage Management Pack from Blue Medora. After logging into the vROps appliance go to the configuration directory for the adapter instance:
Next, let’s open up the configuration file in a text editor:
At the bottom of the file, you’ll find a section which has a listing of the various objects that are collected. (Figure 3)
Figure 3: Adapter Configuration – NetApp Storage
Here we can set the value to “Off”, “Metadata”, or “Metadata and Performance data”. By default, most of the resources will have both “Metadata and Performance data” selected. If you’d like to see the resources and their relationships but reduce the impact of the performance data, then you can set this to “Metadata”. If it’s a resource you could care less about, then just set it to “Off”. Once done, save and exit. You’ll need to repeat this step for any vROps cluster node or remote collector on which this adapter instance will be running.
The third and more surgical method of configuring metrics for collection and analytics would be to leverage vRealize Operations Policies. Log into vRealize Operations as an administrator user and then navigate to Administration -> Policies. Next, select the Policy Library tab. Find the policy which you’d like to edit, select it, then click on the edit button. For this example, we’ll just modify our default policy. (Figure 4)
Figure 4: Editing the Default Policy
Take a deep breath. Policies are very powerful means of dialing in your vRealize Operations environment. At first, they are really overwhelming. Once you start playing around with them, they’ll become second nature.
The first step it to find the specific resource for which you would like to restrict metric collection or analytics. In this example, I’ll use the NetApp Storage Management Pack again. Let’s limit some of the metrics coming back regarding volume performance. Select the “Collect Metrics and Properties” tab in the accordion menu then filter for the object type in question. In this case, we’ll select “NetApp Volume”. (Figure 5)
Figure 5: Editing the NetApp Policy
A list of all metrics collected for NetApp Volume from the Blue Medora NetApp Storage Management Pack will be presented – a total of over 300. The “State” column indicates whether this metric is collected – as either indicated by a grey checkmark showing it is inherited, or a green checkmark showing it is enabled in this local policy, or with a circle with a line through it indicating this metric is disabled. Select the metrics you’d like to disable – in our example we chose all the CIFS performance metrics – select the Disabled option, and then click save. (Figure 6)
You can additionally configure analytics for metrics in the policy. The KPI setting indicates whether this metric is a key performance indicator for this resource. If a violation against a KPI occurs, vRealize Operations Manager generates an alert. The DT setting indicates whether vRealize Operations should calculate dynamic thresholds for this metric. Disabling one or both of these settings will reduce the analytics load on your vRealize Operations cluster.
As you can see, there are a lot of dials and switches that you can use to tune vRealize Operations. Some other articles of interest if you want to learn more about tuning and configuring vRealize Operations are: