VM Slowdown: Volume 1

With the recent release of Virtualization Manager 6.0 (VMAN) and the integration with SAM/NPM there is now a really cool way to view the different levels of performance analysis from application to VM to datastore. By having the data from virtualization manager presented in SAM it provides a single pane of glass for monitoring and troubleshooting the full OS to application to VM to storage environment. Customers often ask me, “Ok so I have some VM’s that are not running well, I think it’s the storage but can’t see this?”

SAM to VMAN Integration

To start with we can monitor the server in SAM, by having the insight at the OS level we can assign out of the box application templates such as SQL, Exchange, IIS, or create custom one’s depending on what we want to monitor. This will then give really deep insight in to how the applications are performing or maybe not performing.

When the virtualization manager integration is enabled with SAM we will be able to see alerts from VMAN that may exist for this server. We also get a storage tab on the server where we can get deep insight to the underlying storage information from VMAN being displayed within SAM.

1.png2.png

Once on this page I can see really specific metrics regarding IOPS and latency, On this particular one there is an alert from VMAN for disk latency

3.png

If I click on “View Performance” and select the latency metric I will be shown the results in the cool performance analyzer view in VMAN

4.png

So what’s causing this latency?

If we close the chart on the top right, it will take us back into the storage tab of this server in SAM. Here we can see what datastore this VM resides on.

5.png

If I select the datastore I can see what applications are currently running. I can also see there are quite a few VM’s, too many VM’s on one datastore can cause read/write latency.

6.png

On the same page I can see the overall datastore latency and the Top VM’s. Quickly I can identify the top VM’s that have latency, with this information I can now decide what to do, this could be to migrate the top VMs off to a separate less busy datastore or simply to migrate them to a datastore running on faster disks. Whatever I decide is the next step I now have deep insight to make the correct decisions.

7.png

This is just one of the queries we get from customers looking to solve issues in their virtualization environment. To read more about the challenges SysAdmins face and how SolarWinds can help solve them, read the eBook “A Day in the Life of a SysAdmin” now.

Thwack - Symbolize TM, R, and C