The resources needed for vRealize Operations depend on how large of an environment you expect to monitor and analyze, how many metrics you plan to collect, and how long you need to store the data.

It is difficult to broadly predict the CPU, memory, and disk requirements that will meet the needs of a particular environment. There are many variables, such as the number and type of objects collected, which includes the number and type of adapters installed, the presence of HA, the duration of data retention, and the quantity of specific data points of interest, such as symptoms, changes, and so on.

VMware expects vRealize Operations sizing information to evolve, and maintains Knowledge Base articles so that sizing calculations can be adjusted to adapt to usage data and changes in versions of vRealize Operations.

Knowledge Base article 2093783

The Knowledge Base articles include overall maximums, plus spreadsheet calculators in which you enter the number of objects and metrics that you expect to monitor. To obtain the numbers, some users take the following high-level approach, which uses vRealize Operations itself.

  1. Review this guide to understand how to deploy and configure a vRealize Operations node.
  2. Deploy a temporary vRealize Operations node.
  3. Configure one or more adapters, and allow the temporary node to collect overnight.
  4. Access the Cluster Management page on the temporary node.
  5. Using the Adapter Instances list in the lower portion of the display as a reference, enter object and metric totals of the different adapter types into the appropriate sizing spreadsheet from Knowledge Base article 2093783.
  6. Deploy the vRealize Operations cluster based on the spreadsheet sizing recommendation. You can build the cluster by adding resources and data nodes to the temporary node or by starting over.

If you have a large number of adapters, you might need to reset and repeat the process on the temporary node until you have all the totals you need. The temporary node will not have enough capacity to simultaneously run every connection from a large enterprise.

Another approach to sizing is through self monitoring. Deploy the cluster based on your best estimate, but create an alert for when capacity falls below a threshold, one that allows enough time to add nodes or disk to the cluster. You also have the option to create an email notification when thresholds are passed.

During internal testing, a single-node vApp deployment of vRealize Operations that monitored 8,000 virtual machines ran out of disk storage within one week.