The success of VMware as the market leading virtualization platform, has created numerous opportunities for third party software vendors to provide tools that monitor the performance, capacity and availability of either the virtualÂ infrastructureÂ itself or the applications that run on that infrastructure.
These tools fall roughly into the following three categories:
- Resource and Availability Monitoring Tools that primarily consume the VMware vCenter data, store it, trend it, report on it and alert on it. This was the first need that most organizations felt when their VMware environments grew and therefore this need was the first one to get addressed (with the large number of vendors in the lower right hand box below). These tools are applications independent in the sense that they do not care what kind of application is running in the guest VM’s.
- Infrastructure Performance Management Tools that also collect the VMware vCenter data (these tools are a superset of the above category) but also add value in terms of being able to either calculate or measureÂ InfrastructureÂ Response Time. IRT the key metric that should be used to assess whether or not the infrastructure is providing the required level of service and responsiveness to the applications, as inferring this from resource utilization patterns does not work in a dynamic virtualized environment.
- Applications Performance Management Tools that collect Application Response Time metrics, and that then trade off depth of root cause analysis in code for breadth of applications coverage.
What emerges from the table above is an interesting picture in terms of the tools that enterprises should choose to manage their environments. The following considerations are recommended in terms of these choicesL
- Understanding how the resources are used in a virtual environment is a critical taskÂ necessaryÂ in order to prevent temporary and growth related resource constraints from impacting applications performance. There are many fine tools to choose from in this regard. The market leaders in this category in terms of market share are respectively VizionCore vFoglight, and Veeam Monitor.
- In order to truly understand whether or not the virtual infrastructure is providing the right level of service to business critical and performance critical applications, you have to go beyond trying to infer infrastructure performance from resource utilization metrics and patterns. A direct understanding of Infrastructure Response Time is required. These tools were profiled in detail in this post.
- Virtualization Aware Applications Performance Management is required in order for the applications teams that support business critical and performance critical applications on a virtual infrastructure to understand the actual performance of their systems as well as to be able to troubleshoot issues. In this category there is a clear tradeoff to be made between depth of root cause analysis (into code) and breadth of applications support. BlueStripe is the only vendor that can provide an end-to-end ART number for any TCP/IP based application across any mixture of physical and virtual infrastructures. However the cost of this breadth of support is that due to the diversity of applications supported, BlueStripe cannot do code level root cause analysis as can products that are based upon byte code instrumentation like AppDynamics, New Relic and Coradiant.
Enterprises who are going to support business critical and performance critical applications on a virtual infrastructure should at the minimum address two needs. The first is to get a true and complete picture of Infrastructure Performance based upon Infrastructure Response Time. The second is to put in place the tools required to monitor these applications in production. A comparison of the newer tools in this space from a cloud perspective was made in this post. It is important to recognize that in the case of both sets of tools, virtualization has created a new set of requirements that are best met by these newer vendors highlighted in this article. Many of the legacy infrastructure and applications performance management tools still needÂ significantÂ updating in order to be able to adapt to the dynamic world of virtualization and the cloud.