How to Track VMs with Vityl Capacity Management
Vityl Capacity Management has performance monitoring tools that enable IT managers to distill information at each level of IT: clusters, hosts, virtual machines (VMs), guests, and minute process details.
Organizations invest significant resources into their IT infrastructures, and you can be sure they want to see some serious bang for their buck. In theory, virtual environments — and their dynamic allocation of resources — offer organizations an optimized, cost-effective sweet spot. In practice, however, things don’t always go as planned.
In fact, common features of popular virtual environments (VMware being the hypothetical use-case hypervisor) like dynamic thresholding and VMotion can unintentionally obscure or normalize inefficiencies. Often, small hiccups turn into “frog in the boiling pot” problems, where incremental changes go unnoticed until long after it’s too late to do anything about them. Other issues, such a lack of longer-term process data, also represent obstacles to efficient use of virtual servers and applications.
To tackle these problems, IT managers need to be able to track each individual level of a VM environment from the top down: vCenter, clusters, hosts, VMs, guests, and process-level details. In all reality, virtualized environments can be incredibly cost-effective, as long as you cover your bases comprehensively.
Keeping Tabs on Small Changes
A common VMware feature is dynamic thresholding, which automatically adjusts resource allocation according to day-to-day use. The thresholds it sets tend to not recognize slowly growing problems. For example, if your VM cluster has a small memory leak on Monday, the dynamic threshold may recognize that leak as the “new normal” on Tuesday. This goalpost shifting can happen over weeks and months, and the manager only finds out when they start to run out of memory and start to page out to disk.
The solution to this issue is to set hard thresholds based on the optimal configuration of your environment. That way, your VM architecture can respond dynamically on a day-to-day basis, without ever exceeding your core capacity limits (and sending you an alert if they do). However, before you can create a running set of KPI watchlists, you need to know what your baseline limits should be in the first place.
Unmasking Hidden Capacity Issues
A second problem is that, because VMware automatically moves VMs around the environment to optimize capacity, some of those VMs never stop moving. Instead of maximizing efficiency, this sometimes obscures inefficient applications.
For instance, CPU overuse, which ordinarily shows up as a persistent, ugly line on a capacity chart, instead looks like daily or hourly fluctuations across many different server hosts. If your overall CPU usage seems normal, you would never notice these “hot” applications. Even if you did, VMware process data disappears rather quickly, often in a matter of hours — if you uncover the issue days later, it’s almost impossible to go back and locate its source.
To reliably identify and resolve this issue, IT managers need to be able to both chart individual, long-term VM motion across the cluster and drill down into process-level details. Often, the causes of inefficient applications are complex — it’s not enough to know that a VM is moving around. If you can point to a particular guest running a specific process, however, you’re in business.
Vityl Provides Automatic Control
Maintaining a virtual environment both has the benefit and drawback of being exceedingly complex. How can VM users hope to know which metrics are important? Which KPIs are needed for what operating systems? Once those are determined, how should those metrics be charted against one another?
Vityl Capacity Management automatically distills that information from a broad spectrum of relevant sources and presents them as need-to-know trends in a simple, easy-to-use interface. IT managers can get a bird’s eye view of the virtual environment before drilling down to granular details — automated analytics reveal subtle trends that would otherwise be missed.
Like any IT machine, VMs present new problems just as they solve traditional ones. And considering the massive size of many virtual clusters, they’re often harder to find. But with the right monitoring system in hand, organizations can confidently wade through the details, unmask hidden inefficiencies, and get the most out of their VM investments.