What is our primary use case?
From an operational perspective, monitoring systems. We've primarily deployed it in production to give better awareness to application support teams of how their apps have been doing. We do have it deployed in pre-production as well, and are actively pursuing these cases in that space as well. It's just easier to sell to executives from the operations side first.
How has it helped my organization?
Mean time to resolution is probably the biggest improvement. Also, operational awareness is probably one I haven't touched on too much. I think in the past a lot of teams really didn't have a comfortable feel for how well, or not well, their applications were performing. You would get log files, things like that, just a small amount of perspective. In a lot of cases, people complaining was their first indicator of a problem and, really, nobody wants to be in that situation.
We started using this product about five years ago, and with AppMon. That was one of our first big wins, getting these big TV dashboards up all over the place. It really gave executives confidence that we did have things under control. At least we knew what was going on. It wasn't always the green check-mark up there, but we knew when there was a problem, before people called in and told us there was one. So that was a huge benefit.
What is most valuable?
- End-to-end visibility of the applications, since we have them instrumented. Understanding where the hotspots are in the applications.
- Quick Fault Domain Isolations: When people are having a problem, understanding where to look, and letting all the other people go back to their day jobs.
- The UEM feature, User Experience Management: Understanding how users are perceiving the application and then connecting that back into back-end systems to understand why things have gotten slow and then dealing with things.
What needs improvement?
Over the last year, one of the things we've had a challenge with is, we've used AppMon for so long, people are quite comfortable with it. We had a pilot of Dynatrace SaaS up and running for about a year, trying to transition application teams over to it: "Try this new tool out," - especially at microservices-based applications. And a lot of the features at the time that were in AppMon were not yet available, or worked differently, so we had some challenges in internally selling it. Since then, a lot of those features have been added into the new platform, and I'd say in some cases have leapfrogged over where AppMon is.
The one thing people really liked that we can do in AppMon is executive dashboards and, until recently, you couldn't even create the business transactions you need for the data at the back-end of those dashboards. But if I had to ask for one thing it would be: In the new platform, give me the ability to do dashboarding in the way it's done in AppMon, and I think that would bridge a lot of the missing pieces. If I could do the same type of executive dashboards on the walls, they would be happy.
What do I think about the stability of the solution?
As far as the new tool goes, we haven't had any issues with stability at all. I'm really impressed with the way they've architected it and made it much more scalable than the AppMon solution.
We've had some challenges with AppMon in the past, mostly due to the top-end Dynatrace server not having an HA solution, which they're addressing now with the new version. So that's a big win for us as well.
These things, when they first started out, were kind of neat and cool but they were never considered mission critical early on. Fast forward a few years to now, and nobody can afford the Dynatrace solution not being online. People are using it to support their apps. It is considered mission critical. So having HA built in, making this stuff rock solid, is very important to us.
What do I think about the scalability of the solution?
So as far as AppMon, we knew where there were ceilings in the product. We've got a pretty big company and we hit some of those scalability limitations early on, to the point we were running about eight AppMon servers just for production load alone. One would never handle it.
The thing that impressed me about the new Dynatrace Managed offering is that people are seeing, here at the Perfrom 2018 conference already, about 100,000 host limitation, at this point. But that's more than enough for us to handle our entire production infrastructure, so I'm looking forward to actually collapsing all of the AppMon stuff into one big solution where I can see, end-to-end, any app offering. Even if it crosses lines of business, you get full visibility. So I'm very impressed.
How are customer service and technical support?
I wouldn't say we need to call them that often, the solution does not require a lot of hand-holding from that perspective, but we've had it for five years so there are always a few bumps in the road. We've had to call in for a few things, and they're really very professional. If we've needed to escalate issues, we've never had a problem doing that. Always been able to get to root cause, and in every case, been satisfied with the outcomes.
Which solution did I use previously and why did I switch?
Five years ago, Wily Introscope was in place for a lot of our WebSphere infrastructure. At the time, we were actually looking at it more as an infrastructure tool. The guys supporting WAS, whether it be in Windows, AIX, Linux, or z/OS, needed something to give them a lot more visibility into what was going on, and Wily at the time just wasn't doing it.
Wily Introscope was only lightly used at the time, but it just wasn't giving them the depth they needed. Realistically, nobody was using it. We had a product that was sitting there, unused. At the time there really was no APM-type thinking at all. It's kind of by chance since we moved to Dynatrace to satisfy that first requirement.
Since then, we've flipped it on its head, and we use the tool that much more from the app perspective. And that's really the way to do it. And we had no problem making that shift at all. We could still satisfy what the WAS team needed to do with the tool but get a lot more value out of it by giving app teams a lot more perspective. We've been able to get 500 percent more value out of it just by putting it with app teams as well.
One of the challenges, when we were looking through all the different solutions out there was, which one supported all of those things to give you that big picture? And Dynatrace met the mark.
How was the initial setup?
I was only loosely involved in the initial setup. At the time we brought Dynatrace in, it was one of my peers in the same larger group that was doing the evaluation. I was in the background, watching it. I then ended up inheriting the whole thing about a year later, so it was kind of good that I did have some visibility into it.
Overall, the initial setup was pretty straightforward.
Which other solutions did I evaluate?
We did a bake-off against some of the competitors at the time, and it was pretty clear Dynatrace was the best fit for our organization.
I know AppDynamics was one of the other ones we were evaluating, and IBM's offering, ITCAM. And given the different types of technologies we had to deal with, specifically mainframe, Dynatrace really stuck out as the one tool that could handle all the requirements we had at the time. And since then we've actually grown the use cases up quite significantly, and are even more happy with it.
What other advice do I have?
When it comes to the nature of digital complexity, the role of AI - when it comes to IT's ability to scale in the cloud and manage performance properly - is huge. We saw it when moving to Dynatrace AppMon and started the Gen 2 platform and wondering how did we deal with problems just a few years ago without tools like this? Now we've gone into the Gen 3 platform space and really, with apps exploding from the perspective of complexity - as it's impossible to manually understand where all the moving parts are at any one time, where to look for a problem if it happens - AI really rolls all that up in a way that makes it useful, and we don't waste time trying to dig through things like we would have even with AppMon. So it's immensely important.
We still use siloed monitoring tools to some extent, and we've used quite a few of them. They have their place for what they do but, in reality, we've had to look at other aggregation-type tools and bringing feeds from each of those together. Let's say a Windows Server monitoring tool can tell you the CPU is higher, but it doesn't give you any perspective on what the actual impact is for the user of the app that's sitting on it. And that one server might only be a small cog in the wheel of the overall app. It's great at telling you what it does, but it doesn't go any further than that. You really need something that aggregates all those things together and has perspective. And that's really what app teams want. I think those tools are better served in the operations space, but then that's where they're stuck, they'll never go further than that.
If there was just one solution that could provide some real answers, as opposed to just data, the benefit would be that looking in multiple tools definitely slows you down. The fact that a tool like Dynatrace can bring all those things together and really help pop problems up right into your face is a huge time-saver. I don't know anyone that's got extra time on their hands at this point, so anything that helps us save time and get us to a resolution of a problem faster is of huge importance. Looking back, a few years back, trying to look at multiple tools to help us figure out problems, teams could spin for weeks because we'd have no idea where to look; classic war-room scenario, where everyone's pointing fingers at everyone else in the room, "It's not my problem." Dynatrace has really done a lot to help us get rid of those types of scenarios and be productive, really addressing actual issues.
I give it a nine out of 10. Compared to the other solutions out there, I'd still say it's top breed. It's great to see they're doing a lot more with the product on a big scale. But there are a few things they need to knock off the list to hit that 10; it's not perfect, but it's really, really good.
In terms of advice, the one thing that I've seen from other app teams internally: Do a PoC. It's much better to see your data in the tool than trying to demo it with easy travel or someone else's app. You just don't get the perspective of what they're looking at is some cases. Set it up, kick the tires, and give it a try, and that will win most people over.