VMware Aria Operations Reviews

The analytics engine:

• Goes through all the data which we are out to collect and gives us a nice dashboard.
• Gives details about the whole infrastructure.
• Shows how things are working now.

We used to do all the calculations manually. Now everything is done within the analytics engine itself. So, it gives us a very clear vision, or data, about what we need in our infrastructure support applications.

A feature that would definitely help is something that automatically resizes the virtual infrastructure. That is one thing that we would like to see in vROps. It would understand the workloads' requirements and then resize them based upon orchestration and automation rules.

I've never had an issue with it.

I've never had an issue with it.

I have never used it.

No. We need the solution because we have to monitor our environment. So for monitoring the virtual infrastructure, it is the key component.

It is a straightforward architecture, but you can make it complex based upon your architecture. Based upon where and how you want to deploy.

All the architecture is based upon your datacenter layout. It's a pretty straightforward process but you have to consider your layout.

In-house.

The most important criteria is that the solution must understand the overall software, defined as it is in your architecture; not only the hypervisor layer but the VMs run on it, the storage, the computing that will be running, and the network as well.

Research integration with any other portal that you want to monitor along with your hypervisor.

I would highly recommend going forward with this solution.

The most valuable feature is the ability to get a view into all of my clusters from each one of my data centers at the same time.

It allows us to meet the needs of our customers a little better as far as capacity management. We can keep ahead of the curve of adding new hardware at the cluster side to meet the demand for what the other teams – such as the Windows team and Linux teams – need for their VMs.

The biggest room for improvement that I can see on vROps is for slightly better definitions around the metrics inside of the reporting. For some of that, it's difficult to find any kind of documentation that explains exactly what you're getting out of each particular metric.

I haven't seen any issues with stability at all. It works great.

Once again, I don't see any issues with scalability.

I have not had to use technical support.

We weren't using anything previously. Of course, we had a few talks with our reps at VMware, and they suggested we use it. We looked into it a little bit and it looked like it'd be a good tool for us.

The most important criteria when selecting a vendor such as VMware is obviously market space and how popular they are. What everybody else's experience with them are.

I was not involved in the initial setup. I heard it was straightforward for us.

I don't believe we were considering any other solutions at the time.

Definitely give vROps a shot; at least if you can get a PoC to see if it works and if it's the right solution for you.

I think it works good. I'm not as experienced with it as some of the other people on my team, so they might have a few more things they dislike or like more about it. But so far from what I've seen, it works great.

I think the best thing I like about vROps is that all the information is in vCenter’s database, but most customers won't take the time to mine the data in vCenter. vROps presents that data in an easy-to-consume format. You don't have to dig in to all the numbers as to why something needs to change or why you might have to improve performance.

It's huge in the performance management arena because it has the ability to really show us that we are over-provisioning a lot of the virtual machines to the point where they're hurting themselves; too many virtual processors, too much memory. The virtual machines can actually be built smaller and perform better.

vROps has the ability to provide insight, fully look at and monitor your environment, and give you recommendations on optimization, efficiency, and risk management.

For instance, let's say that I have a virtual machine that appears to be starving for memory. vROps has the ability to monitor that virtual machine in real time, and give you a recommendation on how to make that virtual machine perform better. Possibly by, say, moving it to another host.

For troubleshooting, it can also be pretty cool, and give you an idea if you're having issues, say, at storage level. Let's say latency has gone too high on one of my LUNs; it has the ability to monitor that by working with SIOC. and some other products on the hypervisor. Also, it has the ability to look at troubleshooting from a performance standpoint; troubleshooting between networking devices.

We have also used capacity management to definitely save on the compute side.

Given that I help others learn how to use the product, in most cases, even with the default dashboards that come with the product, most customers get a wealth of information. Then you couple that with the ability to customize those dashboards for their specific environment. One of the things I've always enjoyed is that we've been teaching and preaching for years on the training side: Right-size your virtual machines to get the best ROIs. The efficiency badge in vROps will tell you exactly how many virtual processors you can reclaim, and how much memory you can reclaim. It's not just an administrator saying, "Hey your virtual machine could be smaller." It's the product actually telling you that you could get a lot more out of your environment.

I would probably like to see better recommendations. I think sometimes the recommendations for performance optimizations tend to be a little too simplistic. The recommendations could be a little bit more in depth, as to why you need to do this or that.

For instance, in a performance optimization, it might say you just need to move this virtual machine from here to here. If you really look at the virtual machine’s overall performance, moving it might be a way to fix it, but resizing the virtual machine might by a better recommendation. Or moving it to another data store might be a better recommendation. I think the recommendations could be a little more tightened up and probably a little bit more in depth.

Also, one of the things I like about vROps is the ability to add the additional adapters to monitor other kinds of products, whether it's NSX, or storage, or even physical hardware to a degree; cloud-based environments. I'd like to see more on that front, to continue developing those additional adapters, those additional third-party add-ons. For example, working with Palo Alto Networks, working with some of our additional storage vendors. There are some good adapters out there, for sure, but there's new stuff coming out all the time.

Finally, I think its ability to interact with vRealize Automation could be enhanced.

We've been teaching on it now for four or five years now.

So far, stability's been pretty good. I can't think of anything off the top of my head where we've had any issues.

Scaling is pretty good, especially when you have the ability to put nodes in remote offices to grab the information then pull it over for it to be crunched. From a scalability standpoint, pretty good.

I personally have not used technical support.

The initial setup is actually quite easy because you just go through the wizard and you deploy your different nodes: your master and your slave secondary, and then your data nodes. It's pretty easy to deploy these days, and then they've got different versions. They've got a version that will run on Red Hat; I've got one customer that uses that. I've got most of my customers using the Windows version.

Definitely get training on it. Of course I represent that, but it's a very complex product. Out of the box, you can get a lot out of it, but there's so much more customization available with vROps. If you'll take the time, you'll get much more out of the product, but it usually requires a bit of training. It's probably taken me a good part of the year to understand a lot of the ins and outs to the product.

We were yet to fully deploy it as it should be. We've yet to fully implement it. Mainly because we have almost the entire HP OpenView suite, so that's been saving our life for right now, but we've noticed the benefits. We've noticed the night-and-day difference from monitoring everything from the hypervisor into the VM as opposed to from the VM's point of view out, without the machine knowing that it's really virtualized, that it’s sharing resources. It believes it is in a physical environment and it owns direct access to the CPU and memory, but it doesn’t: it’s virtual, it’s sharing it and it's pretend. vROps takes that into account because it all goes through the API for VMware.

Definitely, the first win for us was being able to see a CPU wait time where people were still building VMs as if they were building virtual servers, requesting too many CPUs. Not enough memory and they were really shooting themselves in the foot because you can have as many CPUs as you want, but if you're not really using them all, you're sitting there waiting for those virtual slots to fill up before it actually goes to the physical CPU. So you're adding so much overhead.

There were a couple of application teams that were able to take the constructive criticism, per se, and brought down how many virtual CPUs they had and they noticed a huge performance gain. Being able to do that for the environment was a quick win for us.

When you migrate from vCOPS to vROps, it has this awesome API where it grabs all the data, everything you've collected, and it puts it into vROps and you don't really lose that much. Everything you've already collected gets moved over and copied over and you're good to go. However, if you are on vROps and you're migrating to a major version of vROps or a new architecture design – like we're trying to do because we're trying to size it correctly – it doesn't go from vROps to vROps. I believe they had mentioned they were going to do that in the later version, or try to, but that would be my biggest request, because we need to build it out correctly and then migrate all that data we've already collected for so many years.

Aside from that, I would say getting around, creating your own custom super-metrics and all of that: It might not be that it needs to be easier to do, but maybe more well-documented.

Definitely reporting is nice and maybe they could develop an easier UI to do your own custom reports. We're still using all of the out-of-the-box reports, which are great. They've helped us hit that 70% of requirements, but it would be nice to have a nicer UI. Hopefully something like HTML that I can just drag and drop and just play around as opposed to the current UI that I have, which is like a popup; you have to know the metric name, and then somehow click over and get the metric. You really have to know how they're doing it and what they call their metrics and what they call the groups of their metrics and all that to know how to do the report right.

We actually got it back when it was vCOPS 5.2, so I would say we’ve been using it for about three years.

Unfortunately, it has not been a consistently stable solution, because we've never fully deployed it as it's supposed to be. If you go through their sizing guide, we need I believe three virtual appliances tiered and we're currently on one virtual appliance. We have to reboot it often and it's just because it's not sized correctly. That's on us. We haven't had the time. We haven't had the resources. It is a big appliance, one of the bigger appliances that we own, but it's mainly because of what we're monitoring. We're monitoring so many VMs, so many data stores, so many network paths, and all that goes into, I believe, VMware’s equation for how it should be sized.

I believe it will meet the company's needs going forward once we size it correctly. Definitely its internal high availability is very simple to configure. We haven't looked at the disaster recovery for it. Unfortunately, we haven't given it the love that it needs to get it up to the way it's supposed to be, but I believe it will meet the company's needs.

Technical support is excellent; never had an issue with support with VMware.

Try it out. Yeah. Just spin up.

We have access to the software. I don't know how easy it is for somebody else that doesn't have an account with VMware or doesn't have an existing contract with VMware, to get the software, but for me, my solution, for everything that I have questions about, spin it out. They're all virtuals; why not? Worst-case scenario, you erase it. Move on to the next one.

One of the most beneficial features is a single pane of glass as well as the interoperability that it comes with additional features. When you add in management packs, you have the ability to really extend upon the infrastructure and really fine tune what you want to see within the environment. When it first came out many, many moons ago, it was something that was very default and it was a very good attempt. But as it's progressed over the years, it's gotten a lot more extensible so that you can actually fine tune which workloads you want to monitor. They've progressed, and while there are definitely still a few gaps, these are probably the most valuable features.

It helps us do a lot of the VMware optimization assessments, which is one of the first things that our company does. We go in and we understand what the environment is looking like from a one- to seven- to 30-day window, and where their gaps are currently located, so we can deploy an evaluation license of VMware Operations Manager. We can deploy a couple of custom reports and provide them a path forward.

vROps has also helped us avoid outages and shorten our outage time. When you do data center consolidation, you understand a little bit more about the systems in there so you can map where they're going to go.

As far as capacity management, on the one hand, it has helped us, but on the other, it hasn’t. One of the shortfalls with the algorithms of vRealize Operations is that if you want to add additional capacity and you're moving from one, just as an example, Intel generation to another generation, generally speaking, vROps only does it by megahertz, and not the new spec end values, as other tools can. When it comes to capacity planning, if you're doing a consolidation of virtual machines, it makes sense because you're going into a target platform. But if you're adding additional hosts that are a new generation, it's not as accurate.

Similarly, for performance management, it has only partially helped us. There aren’t many vendors that do anything better than CPU monitoring of a virtual machine, because that's what their bread and butter is. But when it comes to memory management, they look at a best guess unless you have an agent in there. That - when you deal with business-critical applications - is a very touchy subject. The perfect example is, you oversize for these VMs, vROps is number one on the list and they say don't resize vROps because it keeps memory for the database. Well, it's the same thing with SQL and some other items, so that's one of the things.

Things that are already on the road map include improvements in guest monitoring when it comes to memory management. Right now, it does a very good guest with VMware tools.

One of the things they'd like to push is Hyperic. Unfortunately, it’s not there yet.
One of the other things that I could foresee happening is, to download different management packs, have it as a link within the appliance rather than having to go to a solution change.

Stability is much improved, much improved.

Scalability is also much improved.

I do not use technical support often. Usually, it's mostly just myself and whoever helps deploy them. So, it's usually the partners that help me out.

I probably have not used it at all, because usually the issues that I come up with, they haven't had a documented issue yet.

I did not previously use a different solution. I saw the demo many years ago and decided this one had some pretty good potential and just followed through with it.
No other vendors were really on my short list at the time.

I'm usually involved in all the initial setups that we do for customers; just not the tool itself. I set the tool up for customers. It's not complex to deploy. Complexity comes with the fine tuning of the policies and what you want to monitor, so you see what you need to see and not a plethora of information that you don't need to know. Setup is easy.

Before choosing this product, I did not evaluate other options.

When I select a vendor like VMware, the decision comes down to customer requirements, and that's what we go from, from the beginning. We go in, we discover, we understand what the customer’s requirements are, and then we use that as a basis of what meets their business needs.

Reporting is product’s the most valuable feature: VM reporting, ESX reporting, trends, seeing where processes could be improved, seeing where resources could be reclaimed; basically, managing and balancing the vCenters. For what I do, most of the things I monitor are VMs. Application owners at times will have problems with their applications running on the VM and they want to see trends such as when are peak times when their resources are being used, such as CP use or memory, IOs and information like that. I could easily give them feedback - run a customized report - and say, "These are your peak hours, peak times."

It gives you a quicker view into the organization, instead of having to go through manual steps to figure that out; it's right there in front of you. We have app owners that use information. Most of the users currently are other IT folks in a virtualization environment.

The application monitoring helps us avoid outages. For example, we set up an application and we monitored it. Then, once the application started using a lot of resources, we're able to realize that we needed to add more resources to this application to avoid any outage.

Capacity management is awesome, too.

Maybe improve the user interface, immediate access to data, and make it easier to get reports, from the application owner's perspective, not for me. I can go in and I can manipulate the data and get what I want. I'm just thinking from the application owner's perspective. They want a quick report to monitor; see a group of VMs and see the process over time. Something like that.

It has been stable; that's good. I haven't had any issues with it. Like I mentioned, even though I do use it, I'm not the main person who manages it.

It's scalable. Right now, I think we have 10 vCenters all over the place. From my perspective, it has not slowed down.

I have not used technical support.

At my previous job, we used VMTurbo. I was looking into vROps on my own. I actually came across it and I saw that it was cool, so I decided to invest in it because of how it was integrated with the virtualized environment.

At my previous job, I started to set up vROps, but then I left. When I arrived at my current job, it was already configured and everything.

It's a great product. It offers a lot of great benefits. It's something that they should definitely look into.

vROps offers a lot more detail that is really helpful for the enterprise. For example, when you're doing performance troubleshooting, or evaluating the efficiency of mission-critical labs, capacity planning, or just looking at environment consolidations, for example, to cut costs. I have actually used it to a fair degree to bring about a few thousand dollars of savings, partially because of how the environment was configured; which is how it should be configured. Those metrics were available through vROps. For example, consolidating the number of leads in a cluster versus adding new clusters for other business needs. There was a level of cost avoidance and there was a level of cost savings at the same time. This was in a previous company that I used to work with that I left just two months back.

We use it to help the team understand how they should be leveraging that infrastructure and how it should be performing. For example, you can talk to your team and ask them to ensure performance at certain milliseconds for IOPS, specific gigahertz for performance, and then you have maybe 60% peak usage or 70% peak usage. There should be capacity for more production workload to go off and run when there's a peak demand in a sudden way; for example, unexpected requirements. At least your environment assigns resources appropriately from that standpoint. If you don't know how your environment if functioning, and you're just relying on real-time metrics, then you're not really planning ahead, and it also can cause a business impact because you don't really know what your environment is doing.

During initial setup, it actually gives a lot of false alarms, so that's one aspect that can be improved, but that's why you have to tweak it to get the right type of metrics.
If it were more agile and more self-descriptive, and in fact, scripted in a way that it just goes and self-installs, and then you specify certain metrics for configuration, that would be awesome.

Also, it needs to catch up with the times. The user interface is really buggy and slow. I'm not sure if it is now on HTML5 or not, but I'm hoping it would be in the latest release. I do not have any experience with version 6 and later. My last experience was at 5.5.

From that standpoint, other improvements would be some intelligence monitoring, and intuitive reporting. Machine learning, if it's integrated with the capabilities of vROps, would be awesome. For example, why should I set an alert at 65% for one environment and 75% at the other, when it might change or fluctuate from time to time. If there's machine learning and it automatically knows the optimum level, that would be awesome. Half of the configuration pain gets cut down right there.

It is rock solid once it's configured properly and it's running.

I have had no issues with scalability, as far as I was concerned. I have actually used vROps for thousands of VMs and have had no issues.

I’ve never had to call VMware for any type of technical support, except maybe for one time when we had an issue with the SQL database - but over a year, one call is nothing.

There is a little bit of complexity initially to set the right metrics and put in certain alerts and tweaks.

There are competitors who are doing something similar, with regards to machine learning. For example, Splunk, not a direct competitor to vROps, but does a lot of stuff that vROps does. CloudPhysics is a great tool. I have also tested out CloudPhysics and worked with one of it in one of the clusters. I can tell you that it’s also a really good product. VMTurbo is another competitor to vROps, but it does a few other things that I might not want to do in an automated fashion in an enterprise. However, there are other things that VMTurbo would be really good at doing where people want that level of automation.

When we spoke to vendors, the detailed metrics was the biggest thing. The level of granularity you get, it's awesome. One bad thing about vROps is the level of granularity from a time perspective; it averages data out at five minutes. If it was possible to go down to 30 seconds, for example, or 10 seconds, that would be really great.

Granularity is good but I want even more, because, to be honest, peaks don't stay around for five minutes. When data gets averaged out at five-minute intervals, you don't catch all of the required information that you need. Still, you get a lot of information out of vROps because you can tweak it for time.

If you have all the core pieces of VMware, vROps is a no-brainer. If you want a little more level of agility, there are other products at play, but it all depends on your requirements. You can't go wrong with vROps. Are there things that are always going to be better? There would be. Would vROps catch up? It would. Would it evolve into something new? It might. Right now, it's there. It does everything you want it to do.

We primarily use this solution to look at our workloads and determine where there's an additional need for space or capacity needed as well as how to optimize it. We've been using it for a few months. 

It has improved the way my company functions in the way that it quickly identifies problems and addresses them. We're still relatively new to it so we haven't used it to the right size. It addresses problems that we see and then we're going to eventually get to the point where reducing capacity where needed will help.

Troubleshooting is one of the most valuable features for us. It identifies problems that other monitoring solutions give us, offers insight into the problem and then digs into it and finds out what the actual problems are and addresses them.

We have found this solution to be intuitive and user-friendly. The dashboards on the home page are easy to navigate. Once we log in, it's easy to find what we're looking for and to see immediate problems.

Stability is good, we haven't had any issues with it. 

The solution is intuitive so I haven't had the need to use their technical support.

I rated this solution a nine because I haven't had any issues with it and it has been intuitive and easy to use. I don't know it well enough to give it a ten.

If you're considering this solution, I would advise you to consider it and to look at your environment to see how it can assist with troubleshooting.