It’s quite common that we find our virtual machines (VMs) running slow, or the apps running on them becoming sluggish. There can be numerous reasons for VM issues, and some of them are caused due to the dependency between the VMs and the other elements in the virtual environment such as the host or datastore. Let’s understand some common cases of VM performance problems.
- A VM could be slow because the host doesn’t have enough resources to allocate to the VM. This is a resource contention issue at the host level. There is VM resource demand, but the host is not able to provision requisite resources.
- A VM could be slow because there are too many VMs accessing a datastore and contending for storage. The datastore faces the situation of catering to more VM load than what it can actually support.
It does become difficult to identify the source of the VM bottleneck even if you have alerts set up to notify you of an anomalous VM metric. What you need here is the relational understanding of which VM is run by which host as part of which cluster.
SolarWinds Virtualization Manager provides an informative interface to view the entire virtual environment in the context of a chosen element (such as a host, a VM, a cluster, or a datastore). When you can get to understand the relational dependency between the various elements of the virtual infrastructure, you will be at a better position to infer the cause of resource contention if that’s what is causing your performance issues for a VM.
For each selected virtual component, you can see the mutually associated other elements of the virtual infrastructure. For e.g., if you have selected a cluster, you can relationally see
- all the hosts that form the cluster
- all the VMs that run on all the hosts in that cluster, and
- all the datastores that attached to all the hosts in that cluster
This works in other possible combinations too. A selected datastore would tell you which VMs are accessing storage from it, and which hosts are attached to it, etc. And likewise track down VMs and clusters.
The best part is, as you see all of the virtual infrastructure from VM to host to storage, you will also be able to see the alerts on each of these elements right on the environment map. Just a mere mouse over action will tell you what issues a VM, or host or datastore is having. You can drill down from there to explore further.
So, if you find a VM which is running slow, you can identify the associated datastore and diagnose how many other VMs are consuming resources from the datastore and where this is causing any resource hogging. With this level of information, troubleshooting becomes extremely simpler. You can decide whether to provision more resource to the host, or load-balance the VMs based on business requirements.
Time Travel is another feature of this dependency mapping which allows you to “go back in time” to see the dependencies that existed at a point in time in the past. For example, you can see all of the VMs that were on a particular datastore 3 days ago and which ones were generating the most I/O.
Watch this short video by Sott Lowe (vExpert) where he explains how you can use the virtual environment mapping and time travel features of SolarWinds Virtualization Manager to contextually understand performance issues in your virtual infrastructure.