Wednesday, March 8, 2017

Using End Point Operations agent with vRealize Operations Manager

End Point Operations as a feature of vRealize Operations Manager is an area where I get a lot of questions. During a workshop with a customer today, I was asked about a number of things which rang a bell in my head. I spent some time with the customer to de-mystify a lot of apprehensions around EPOps and the conversation ended into a meaningful action plan.

The phenomenon of agreement in the IT world (especially operations) is rare and hence I think that it would be worth to share the findings with others who might be looking for similar answers. With the slow departure of Hyperic and End Point Operations becoming mainstream, it is important that customers who have use cases for monitoring beyond the hypervisor layer (which vROps does for you OOTB) have a clear visibility into where End Point Operations stands.

It is important to understand that Operations Management as a process is not limited to a tool or feature. If you walk-in into any medium or small sized customer environments, you will find a number of tools solving point problems. I am not saying that vROps can solve all the problems of the world, but I do see customers inclining towards tools consolidation and considering vROps as the mainstream platform which can help them reduce all the complexity while covering the most number of use cases they would have around Operations Management.

While I can go on and on about Operations, the scope of this post revolves around End Point Operations. The use cases are pretty specific, however they can go DEEP and WIDE as we are moving from the hypervisor layer and into the zone of Operating System, Middleware and Applications. If you have been into IT operations, you would quickly realize the complexity you need to deal with once you look at all of this data in silos using multiple tools. Hence, if a tool can provide capabilities for joining all the dots to give you a complete picture right from the physical hardware, up-to the response times of a transaction one is running on a OLTP platform ALL IN ONE PLACE, it is a easy sell.

The statement I just made is also known as the "HUNKY DORY WORLD". Only if it was this simple, many of us would not have a job :-)

End-Point operations in its current form, and time to come is eventually trying to solve this problem. Post the integration of EPOps solution within the vROps platform, the unified view is no longer a distant dream. Having said that, I have to be brutally honest in saying that this UNIFIED VIEW needs a great amount of expertise to ensure that it is build right. With the evolution of EPOps, I am glad to share that there are a number of solutions which VMware has been able to build and deliver and with this pace of delivery, I am sure there are a number of them on the roadmap (only if I was allowed to share the roadmap without an NDA ;-) ).

I must call out that the list I am about to share are the one's which are made by VMware. Blue Medora, a very close partner of VMware, has delivered a number of management packs and plugins with un-matched quality and they are worth exploring as well.

Here is the list of available vROps plugins for end point operations manager. These will help you monitor the applications and related services.

I would highly recommend that these should be tried in a Test/Dev environments to see if they meet your use cases and then deployed in Production. You can use the links to download the plugins and the related documentation for more details:-

Some best practices around installing management packs & EP Ops:-

  • Ensure that you test them in a Dev environment. This includes, the metrics, the relationships, the new alerts and dashboards which are added by the pack OOTB.
  • Once you have the POC done and content finalized, install the pack in Production.
  • Disable all the OOTB Alerts and enable what you need. Customise them to meet specific use cases. It is always a good practice to disable all the alerts and clone the OOTB alerts (whichever you need). This will not impact the alerts you chose to keep during major product upgrades.
  • Export any unwanted dashboards introduced by the management pack (for a backup) and delete them. This will help avoid clutter.
  • You might want to be selective with your approach with deploying EP Ops as per the sizing done for your vROps deployment. If you plan to go beyond the capability of your vROps cluster, please expand it by using the Sizing calculator.

With this I will close this article. Hopefully the information above would tickle your brain to see how you can deliver the unified view which you always wished for.

Share your thoughts in the comments section and as always, Share the Post.. You never know who needs it!!

No comments:

Post a Comment