What is most valuable?
The performance analytics, when we are troubleshooting performance issues. It's good to have insight into special metrics like CPU ready time, latency, network utilization. Otherwise, you're kind of like shooting in the dark, guessing what the problem could be.
How has it helped my organization?
From a conservative aspect, we can use the data to, for example, for purchase decisions. If someone asks for something and they are not using it, we're taking it back. So from a full-cost perspective or a show-back perspective, we are able to use that data and to say to application teams, "You are spec'd out, way over-provisioned." We're not seeing that so, for the business we are able to reduce cost by purchasing hardware because that data is showing us what needs to be and where it should be. So if application X is over-provisioned but application Y needs the utilization, we can shift those resources.
What needs improvement?
My number one request personally would be self-healing. So if there is an issue with the appliance, we should alerted and it should be clear as day. When you log in you get all these dashboards, everything looks really cool. But those dashboards don't do us any good if the health of the appliance is not 100%. So, if there are health issues with it, or it's not collecting data, it should self-heal. Or if the data is filling up on the disk, we need to know that and be able to click a button and say, "Do something about it," or "Give us step by step instructions on what to do to add a disk."
I think they could also work on the infrastructure a little bit.
It's highly customizable but it's hard to learn it. You have to be in there every day to really get the best use out of it. The nature of our organization is that my team kind is in charge of it but it really shouldn't be my team. It should be a monitoring performance type of team or operations team that owns it so that they can put the time in and create the proper dashboards. At the end of the day, they are going to be looking at those dashboards, not us.
So the ease of use: If you're not a vROps guy from the beginning, it's a high learning curve. It comes out of the box with all these settings, that is where training comes in. I know they offer courses for that. But I think the solution should be more natural in getting to know it.
What do I think about the stability of the solution?
I have some worries stability. Initially it was slow, so we ended up adding resources to it because the environment grows. That solved it. I'm not a big fan of the master data model that it has, so when we have failures, it's not always clear if the collectors are up.
What do I think about the scalability of the solution?
You have to keep an eye on the actual appliances. If you're scaling higher, you're going to have to add add resources to the appliances themselves. Otherwise they slow down.
How are customer service and technical support?
Their VMware technical support is always pretty good. In our specific case, we had a special issue that they couldn't resolve. Overall I would say it is good, when you get a hold of them.
We just open a ticket up on the web, we set the severity, and then they will work with us from there. They make it really easy to open them. I'll do my best to troubleshoot, but I only have so much time in the day. At the end of the day, I'm just going to open a support case and they'll help me directly.
We pay for support so, the business wants us to use it, take advantage. They look at how many support cases we open. Otherwise they might say, "Hey, you guys don't open up support cases. We're paying."
Which solution did I use previously and why did I switch?
No. We just had the foresight to know we needed a solution like this because server sprawl is very concerning to me, personally. I don't want to have a whole bunch of hosts out there that I don't really need, because the bigger you get the harder it is to maintain it.
Which other solutions did I evaluate?
At the time, the other one was Turbonomic, but they are a little higher-end. We had to "sell" it to the business. vROps is not cheap either. We convinced the organization that it is in their best interest. And they followed suit.
What other advice do I have?
There are different things to look for when choosing a vendor. From an engineering standpoint, it's the administration. From an organizational standpoint, it could be cost. So it has to come in between those and it has to be a stable product as well. Those three factors.
As engineers, we're the decision influencers but at the end of the day we are not cutting the check for the organization. So, we have to do our job to sell it to the organization if we think it should be recommended. They have to have total buy-in.
I think the decision depends on your server infrastructure. If you're hyper converged, your solution may already have those analytics built in to it. So first check if your infrastructure or server provider already has that. A Nutanix may have that. If you are a traditional shop with blade or rack and you know you don't have it then there's really no competition besides Turbonomic.