VMware vRealize Operations Manager performance issues when a single VMware vCenter Server Adapter instance collects data for thousands of vSphere objects
search cancel

VMware vRealize Operations Manager performance issues when a single VMware vCenter Server Adapter instance collects data for thousands of vSphere objects

book

Article ID: 342688

calendar_today

Updated On:

Products

VMware Aria Suite

Issue/Introduction

Symptoms:
When you have a single VMware vCenter Server Adapter instance collecting for thousands of vSphere objects and the VMware vRealize Operations Manager cluster is sized appropriately for the number of objects and metrics being collected, you experience these symptoms:
  • The vRealize Operations Manager UI appears slow on the node where the large vCenter Server Adapter instance is located, compared with the other nodes in the cluster.
  • You see this alert generated for the node where the large vCenter Server Adapter instance resides:
vRealize Operations Cluster processes may not have enough memory

1 out of 6 descendant vrealize operations analytics(s) have symptoms vRealize Operations Analytics is experiencing a high amount of stop the world Garbage Collection





Environment

VMware vRealize Operations Manager 6.1.x
VMware vRealize Operations Manager 6.0.x

Cause

This issue occurs because the node responsible for running the large VMware vCenter Server Adapter instance may need additional resources beyond what initial sizing estimates recommend. The vRealize Operations Manager Sizing guidelines operate under the assumption that collection for objects is evenly distributed among the nodes in the cluster. When there is a single large vCenter Server Adapter instance, the performance is impacted.

Resolution

Additional memory can be configured on all nodes in the cluster or a remote collector node can be added to the cluster for running the large VMware vCenter Server Adapter instance. If VMware vRealize Operations Manager 6.1 is in use, multiple remote collector nodes can be configured in a Collector Group to provide resiliency for the large vCenter Server Adapter instance.

Adding a remote collector node (or nodes) is the preferred solution as it results in less wasted memory on the nodes where the large vCenter Adapter instance is not running.


Additional Information

If a remote collector node experiences a failure, adding a single remote collector node in VMware vRealize Operations Manager 6.0.x may result in missed metric collections. With HA enabled and the large vCenter Server Adapter instance running on one of the data nodes, the adapter instance may automatically fail over to another data node in the event of a node or service failure.

For more information, see:


vRealize Operations Manager Sizing Guidelines
单个 VMware vCenter Server 适配器实例收集数千个 vSphere 对象的数据时,出现 VMware vRealize Operations Manager 性能问题