This document is meant to be HOWTO that describes the steps that can used to inject metrics into vCOps by way of a Powershell script that leverages the vCOps HTTPOST adapter. Powershell will serve two distinct purposes for this document. The first is to be leveraged as a data collection method via existing Powershell scripts (will show how to make a basic one) that do the heavy lifting and conversion of data from an end device to a more friendly Powershell Object. The second is to work with the HTTPOST adapter and in a uniform manner post the metrics collected to the vCOps Enterprise console.
The links below include the document as well as the Powershell script that does the work for you.
This is as easy as it gets!
UPDATED for vCOps v5.0!
-When running ps_vcops_httpost.ps1, augment the parameters with "-protocol https -vcopsuser user -vcopspass pass"
Do you have any real or anecdotal information on how VCOps has helped provide insights with a storage environment? Any powerful ROI situations where VCOps added value from a performance management perspective? Reason I ask is whether it is complementary to the deep performance tools any storage admin has access to.
I believe I understand what you're asking here.
It can definitely be complementary to "purpose-built, product specific" monitoring tools that a storage admin may have access to. However, there is some overlap here. Depending on the performance tool used, I believe its main value which differentiates it from other products has to do with understanding core things about storage in order to help model or present data in a meaningful way. This can be response times, RAID penalties, relationships, failure scenarios, etc.
If you are looking at vCOps as a product, it internally as a "solution monitoring" technology does this kind of thing, mainly in analytics, currently focused at the management and hypervisor layers. In solution monitoring what is important just like deep performance tools, is that we understand relationships. With vCOps it understands its solution by a VM living on a hypervisor, which accesses datastores to access its virtual disks.
I believe that the vCOps tool is complimentary as well as a competitor. If the vCOps has the chops internally to setup relationships (it does), do threshold alerts (it does), generate reports (it does). If these are all so then we can start to dig into the benefits of what an analytical engine and a global solutions view comes into play.
At the end of the day where we are driving is towards cloud monitoring solutions which dictates a transactional view. In order to make this sustainable we need something that looks at transactions across any part (global) of a cloud and understands the health of that transaction. It then can look below the covers and uncover abnormalities all the way down to the backend disk. This is what I would dictate the solution monitoring where we are very much relationship minded and doing “top down” analytical review of the dependent stack vs. a bottom up silo and threshold based monitoring strategy.
Real world view of vCOps helping provide storage insights? VMworld 2011 in Las Vegas was the first run with VNX stats being displayed in vCOps. During that show there were performance issues being reported. vCOps was able to identify for us that storage was not the issue and pointed the VMware team to the source of the problem. A quote from the show.. “I previously was driving the Enterprise and asking for more power by yelling to the engine room, ‘more power’! Now with vCOps and bringing all this information together I as the captain am in control.”
Does that mean that there is no vCOPs storage adaptor needed to monitor VNX block metrics like e.g. disk service time and storage processor memory and CPU usage?
That is correct, see the attached powerpoint presentation for details about how it works and what kind of block stats are available with the VNX. This is not to say there will not be further methods to get VNX stats into vCOps. I can only comment on what we have available via the work presented here.
Clinton - this is very cool. I realize this isn't intended for "production use" right now. What is the license on it, and can I modify it for my own lab use? I looked for any license info but didn't find anything. thanks!
This can be used any which way, production/non-production/lab. The big disclaimer is that the use of the scripts is not directly supported by EMC, as what was posted here is meant for lab use cases and accelerating the understanding of metrics and use cases in vCOps. The method we use to get the metrics into vCOps is a VMware 100% method (vCOps open adapter, REST API). The powershell scripts simply leverage this supported vCOps interface and make it a bit easier to work with. This open adapter that we leverage here is actually used by VMware in certain production adapters.
Things that make a lot of sense using this injection method are likely candidates for an official EMC supported plugin (ie. VNX Connector).
Feel free to modify to your own needs and let us know what you're doing with it!
I tried this in vCOps 5.0.1, but got lost when logging into vCOps. There doesnt seem to be an Environment Overview section like you described in yout document.
Any idea where to look for the new information?