vRealize Operation Manager, Who Needs it?

In these years, Virtual Machines appear on top of Physical Hosts, Virtual LUNs appear on top of Physical Storages, Virtual OS and Applications (Okay, Dockers) appear on top of VM/Physical Hosts, even we have build virtual networks over physical ones. There is a trend of building stacks of abstraction layers one over another. If you have or are exporting Cloud Native Application or Containers Architecture like me, you may be dizzy about all the abstraction layers (well… at least I were).

containerarch31

But I am not going to talk about Cloud Native Application stuffs in this block, I mentioned the above because whenever an abstraction layer appears actually I saw most people just then skipped the underneath layer(s). Life can be good when you didn’t abuse or exhaust your underneath layer(s), e.g. you can always thin provision more and more stuffs until the underneath storage used up, but if your underneath storage died… all the above abstraction layer(s) won’t stay alive.

1355761810568

And thus, in this blog, I want to state the increasing importance when we are building and stacking abstraction layers. This is explicit actually, in the ages when we are having physical machine, network, storage, the monitoring is comparatively trivial. When VM age comes, so how to monitor VM abstraction layer over the physical layer becomes important, this is why new solutions like vfoglight, TurboVM…etc. come out. Story continues when one adopts virtualised storage, network, OS and application, you would need corresponding tools for monitoring those too. Definitely, it is good to have one monitoring tool to keep an eye on every abstraction layers and components inside . But this will lead to an explosion in log messages and alerts which could be non-organized and messy. So what will you do mostly? Well… I agree with the approach taken by many of my customers, to:

  1. Drop the monitoring tools
  2. Setup an Email Rule to Delete all monitoring Alerts

So either way, we just ignore any warning and alert messages in the environment. Of course everyone know this is not proper, but anyone being the system administrator or operator receiving thousands of such messages everyday, they will definitely force to do so.

all-the-things

This is why I think how well a monitoring tool is not defined by how many stuffs it can receive message from, but should be how useful every single message is sent out from the tool. Or course, some tools in the field are highly flexible that you can and thus have to do some customisation and development before the tools can give you meaningful alert messages. While some other tools give Out-Of-The-Box Intelligence and extend through Plugins. While vRealize Operation Manager is one of the later tools type.

vRealize Operation Manager (vROPS)

This is a default monitoring tool for vSphere environment recommended by VMware definitely, when people usually compare vROPS with Solarwind, TurboVM, vFoglight or lately with Elastic Search, I think we better understand the nature and objectives of vROPS rather than a technical feature by feature comparison. To make it explicit, from field experience, if you are looking for:

  1. Real time monitoring (I mean <1 minute interval)
  2. Unstructured logs monitoring (You know, Splunk stuffs)

Then, you can skip this blog post and vRealize Operation Manager. Because vROPS is NOT designed to handle to above requirements, you would need some other other tools if you are really interested in the above requirements, say VMware Log Insight for Unstructured Data Analysis. But what vROPS does or what vROPS wants to do for you is the following instead:

  1. Alerts Reduction
  2. Proactive Environment Monitoring
  3. Capacity Planning
  4. Environment Reporting

Well, the BAU tasks above are tedious but also actually consumed most of my customer’s time in their normal daily work. As a system administrator or operator, the most valuable tasks would not be fixing issues or following service requests. Their talents should be working in future project planning and deployment for fulfilling User Requirements which is more revenue driven, which is more valuable. So if you are belonging to one of the categories below, you need to have vROPS helping you out:

  • Category 1: Furious vSphere User, tired of using vSphere Client for Troubleshooting
  • Category 2: Panic vSphere User, scared unexpected issue happening in the environment
  • Category 3: Confused vSphere User, in reading Performance Metric in vSphere Client for Capacity Planning
  • Category 4: Tired vSphere User, in copying and pasting VM information into daily/weekly/monthly Report

matrix-morpheus

Yes, if your are one of the tired, scared or confused user of vSphere, you would need vROPS. It can help offloading and solving a lot of problem you are facing day by day. So instead of listing all the function and feature one by one which you can read from a lot of blogs already, I would like to illustrate how vROPS can help a System Administrator every day.

Give Meaningful Health Status And Alerts

So first of all which quite a daily routine task that a system administrator or operator has to perform would be health checking, I have blogged the Daily Health Check Method using RvTools and vCheck PowerCLI scripts before. But here, we are referring to a more ad-hoc and close to real time health status monitoring.

I don’t know if you are agreeing or not, but most operation people I met, they actually will NOT drill down or deep dive in an environment to dig out problems as a daily routine. I mean actually this is not something expected to be done daily actually. Instead, what they have to learn about is actually something comprehensive yet simple which is an overview of the environment they monitoring. This does not require lot of metrics, messages or logs, this needs in opposite holistic diagram, charts or maps which can let them spot abnormalities and issues in the environment.

Moreover, monitoring pattern also has to change from reactive to proactive in order to prevent something happening and discover problems before your end user telling you, which is always the case. This is why the analytic components in a monitoring tool have to be intelligent enough to discover future issues. By the way, do not think this is a Minority Report feature. Thinking it as stock market analytic tool, we just need some mathematical formula to calculate what the performance metric should be expected and if the actual metric differs from it for too much then it indicates a possible issue. I think stock price tool like those bands working in the same way?

VROPS is one of the tools giving you health status based on the above design principles. This was a unique feature when VROPS was still called VCOPS (vCenter Operation Manager), i.e. long long time ago. So what making VROPS still unique now? Two things:

  1. Large Eco-System in Extendability
  2. Out-Of-The-Box Configuration

The above features making monitoring ridiculously simple, everything is so straight forward enough since initial deployment. So followings are the high level (and detail) steps:

  1. Download the OVA/OVF based VROPS Virtual Appliance from vmware.com
  2. Import it into your vCenter
  3. Power it on
  4. Initial Wizard to setup the vROPS (5 pages in the Wizard in total)
  5. Add the vCenter Entry you want to monitor
  6. Start Monitoring

With the knowledge of VMware Products from VMware, VMware Realize Operation Manager (vROPS) provides you the best recommendations again threshold definitions, resolution suggestions and resource optimisation indices. Such that you don’t have to spend time in defining your own monitoring policies in most of the cases. You don’t have to redefine or reconfigure anything when you are adding another ESXi host into your vCenter, adding or removing a Workload in your environment.

screen-shot-2016-10-28-at-9-58-27-am

Provide End to End Visibility

So what if you also wanna monitor other stuffs like Hardware Storage, Network or Application running in guest OS? As said, VROPS can be ramped up with management packs to cover those, the setup is also intuitively easy. But I would like to state, it’s not enough to monitor those as separate objects, VROPS has the intelligent to link up Physical Stuffs with Virtual Objects and let you know the relationship among them. It gives you a better Visibility of your environment say if you want to know:

  1. Which VM will be affected if there is a storage port failed
  2. What’s the capacity usage of a Thin Provisioned LUN supporting your datastores

Well, these can be done without VROPS, yet very very tedious and could spend you a lot of time to figure out the mapping. But with it, it’s just displayed in a single Dashboard.

screen-shot-2016-10-28-at-9-59-37-am

Scenario Based Capacity Planning

Many Customers of mine have their own Capacity Planning Standard in their Virtual Environment, most of those are quite static. Say they define the CPU and Memory Consolidation Ratio to some fixed 1:4 and 1:1.5 Magic Numbers. Well, technically, this is not good as it could waste a lot of resources and didn’t take care of other metrics say Storage and Network Capacity. To be specific, I mean the Contention when you are adding more and more hosts in the environment without upgrading the SHARED Storage, Fabric and Network. VROPS gives you a comprehensive Capacity overview of all these factors which should actually be considered.

But I know most of the company Capacity Planning Policies just won’t simply changed because of that. Yet, I would recommend you to at least using the Scenario Based Capacity Planning feature in vROPS to visualise how the environment will be improved or impacted when you are doing some Capacity Planning tasks for your new projects which could Add/Reduce VM, ESXi or Storage.

screen-shot-2016-10-28-at-10-03-41-am

Environment Reporting

Last but not the least, another most clumsy task an administrator/operator has to be performed is reporting. As BAU team, we have to create reports for Capacity, Usage, Compliant, Efficiency, Configuration… etc. on and off. This can be done again actually without vROPS, but you need to have some scripting skill set like running and customising the vCheck PowerCLI report or SQL Reporting Skills to directly extracting information from Database to create a Report. However, if these sounds too much effort, then you should again let VROPS help you. Out-Of-The-Box, we have few twenty Reports you can leverage and you will get even more when you are installing 3rd party management packs. If these are really not enough, we provides a very intuitive Report Customisation Wizard to let you create your own Reports.

screen-shot-2016-10-28-at-10-01-39-am

Conclusion

So to align with the Topic of this Blog, vRealize Operation Manager, Who needs it? I think if you are one care about your environment and you want to have a useful monitoring tool back. You would need it. If you are one care about what are you doing and you want to do things efficiently with your edge, you would need it. Let vROPS helps you and gains back your time!

one-does-not-simply

NSX in Real Life
VMware Virtual SAN (vSAN) Day 2 Operation – Scaling your vSAN

Leave a Reply

Your email address will not be published / Required fields are marked *