Multitenant Use of vRealize Operations as a Service : Designing for vRealize Operations as a Service : 4.2 Architectural Overview
   
4.2 Architectural Overview
vRealize Operations Manager is available in two different deployment models—as a preconfigured virtual appliance (vApp) or as a Windows or Linux installable package.
The vRealize Operations Manager vApp encapsulates a VMware virtual machine. Each instance of vRealize Operations Manager can be configured to perform one of the following roles within the complete vRealize Operations Manager instance/cluster:
Master node – First node (mandatory) in a cluster or single standalone node.
Master replica node – An optional node to provide high availability (HA) to the master node.
Data node – Used to scale out a vRealize Operations Manager cluster.
Remote collector node – Performs data collection only, and can be used behind firewalls or across limited bandwidth connections.
The following table lists the components that make up a vRealize Operations Manager node.
Table 2. vRealize Operations Manager Logical Node Architecture
Component
Description
Admin / Product UI server
Web application that serves as both the user interface and the administrative interface.
REST API / Collector
Collects data from all the components in the enterprise.
Controller
Handles data movement between the UI, collector, and analytics.
Analytics
Analytics engine performs data correlation, calculation of metrics and super metrics, capacity planing, and alert generations.
Persistence
Performs database operations across all nodes in the solution.
FSDB
File System Database stores the raw metrics in the file system. Each node of a vRealize Operations Manager cluster contains parts of the FSDB for overall persistence.
xDB (HIS)
Historical Inventory Service data is maintained only on the master node and master replica.
Global xDB
Data that needs to be persisted on each master node and/or replica like dashboards, reports, policy settings, alert rules, and super metric formulas, which can not be distributed across the system.
 
vRealize Operations Manager can be horizontally scaled out to multiple nodes to monitor larger environments and provide high availability. Each vRealize Operations Manager installation can scale to a maximum of eight nodes. There is a minimum of one master node. In a high availability configuration, this is accompanied by a master replica. All other nodes are data nodes. Remote collectors are not counted as cluster nodes.