vCloud Ecosystem Monitoring Architecture

This is a re-post of the VMware blog article I wrote over at

http://blogs.vmware.com/vcloud/2012/07/vcloud-ecosystem-monitoring-with-vcenter-operations-and-vfabric-hyperic.html

At the beginning of this year, a team within VMware looked at how we could utilise vCenter Operations and vFabric Hyperic to create a vCloud Monitoring solution. The following blog post shows how this architecture fits together and is integrated. I wrote a previous article giving an overview of the integration points, which can be read by clicking here.

Architecture

The integration of these products relies on a vCloud environment following the vCAT reference architecture. There are numerous articles written on how to design your vCloud environments, which this article will not go in too, however the following diagram depicts the logical representation of a vCloud Environment as taken from the vCAT documentation.

The Management Cluster contains all the vCloud Infrastructure Virtual Machines that are required to provide the vCloud ecosystem. The following list depicts the minimum components required for a vCloud environment to exist.

• vSphere
• vCenter Server
• vShield Manager
• vCloud Director
• Database server (Oracle or SQL)

These components along with the scalability of vCloud Director must be taken into consideration when deploying the monitoring environment. vCloud Director can provide support for up to 25 vCenter servers.

Integration – Logical overview

The following diagram depicts the logical design for a vCloud monitoring deployment.

As demonstrated in the diagram all the vCloud infrastructure components are monitored either through vFabric Hyperic relaying information to vC Ops, or through vC Ops adapters themselves.

The links between vFabric Hyperic and the specific virtual machines demonstrates that vFabric Hyperic is monitoring the services and processes running on the vCloud Director cells. Once an alert or status change is received this is relayed to VMware vC Ops. Virtual machine status is monitored directly through the vC Ops vCenter adapters. This allows us to monitor the status of the vShield Edge and vShield Manager virtual appliances.

Through the vC Ops vCenter adapters this allows us to monitor the virtual machine status of all cloud workloads. We can determine if these are up or down, however no information regarding the actual state of the application workloads running in these virtual machines can be monitored.