Announcer: This episode of SolarWinds TechPod is brought to you by the SolarWinds APM suite. Simplify your full-stack monitoring for hybrid, Azure, AWS, and other cloud-native IT environments, and gain valuable performance insights across user experience, applications, services, and infrastructure. Visit solarwinds.com/apmsuite.
Patrick: Welcome to SolarWinds TechPod. I’m Patrick Hubbard, your host for today’s episode, and with us today, it’s a pleasure to welcome, Mike Pfeiffer. He’s a 20-year tech industry veteran who’s worked for some of the largest technology companies in the world, including Microsoft and Amazon Web Services. He’s also the host of the CloudSkills.fm podcast.
Patrick: In this episode, we’re going to make sense out of cloud and application performance management, which most folks just call APM, and specifically, we’re going to talk about how to determine which cloud provider might be best for your needs. We’ll talk about some multi-cloud strategies for application delivery. We’ll talk about the rise of custom applications, and ultimately, how tech pros can manage all of this added complexity by taking advantage of APM.
Patrick: Mike, it’s great to have you with us.
Mike: Hey. Thanks, Patrick. Super excited to be here.
Patrick: Yeah. Well, thanks again. You have a really broad background in tech. How did you end up focusing on cloud, and particularly cloud-native?
Mike: Yeah. It’s a good question, and I think there’s a lot of people that are getting to that place now. I was luckily early on cloud, because I was just really paying attention, and I’m not saying that people that haven’t been on cloud are not paying attention, but I’m hyper-focused on what’s happening in the industry because that’s part of my job and what I do.
Mike: So, in about 2010, Microsoft was pushing towards Office 365, and all of my customers were still on-prem. I was really excited about something new, because I love to learn and I love to geek out on what the new stuff is, and then I realized in that timeframe, though, that not everybody was super excited about going to the cloud, but I could kind of see that’s where this was going. When I saw the traction that Exchange online eventually ended up getting, and most of my customers moving that on-prem workload into the cloud, it just started to seem obvious to me that that’s the direction this is going to end up in.
Mike: So, I made the choice in, like, 2012 to kind of switch off of traditional infrastructure and start getting into cloud. I was working at Microsoft in that timeframe, and I ended up going over to AWS to work there. Luckily for me, it was right place, right time, right time in my own personal career, right time in the industry.
Mike: So, I just saw the value there and luckily got in early, and it’s been pretty awesome ever since. I really enjoy it. And that’s kind of how I got started in it.
Patrick: Well, it’s a pretty unique background, because it combines both Microsoft and Amazon, and you don’t typically see that. So, thanks again for being on the show today. Let’s just go ahead and get started.
Patrick: I think the first thing is, you spend a lot of time on the road working with companies and engineering teams that are actively making the move to the cloud, right? They’re not just talking about it. So, where are we sort of… maybe not so much as an industry, but as operators, on that scale of talking about it versus actually moving production workloads?
Mike: Yeah. It’s a really good question, because the first couple of years… So, I left AWS in 2015, at the end of 2015, started my company now, and in that timeframe, we were working with mostly AWS customers because, as most people know, AWS was first, and most people that were going into production and really using it were using that platform.
Mike: So, in that timeframe we were working with a lot of AWS customers, but not huge enterprises, and that’s really my bread and butter. So, the first couple of years was not a lot of enterprises, but the enterprises that were kind of looking into cloud, whether it was Azure or AWS, were just testing. They’re doing very basic work.
Mike: This past year, in 2019, it’s been completely different. The enterprises now that we’re working with are doing real projects and doing more work in Azure than they were in the past couple of years. So, we’re still kind of in the enterprise, it’s still early. There’s still a lot of confusion. There’s still a lot of uncertainty. It’s like I’ve been telling my friends, it’s the Wild West out there when you’re working with enterprises, just because there’s so much new stuff. It’s not easy to take a traditional application, whether you’re lifting and shifting and moving it over, or trying to transform it into something else and use some kind of bleeding edge technology. It’s just a lot that goes into it.
Mike: I think probably one of the biggest things that enterprises are having a hard time with is just getting started with the governance piece, the security, figuring out the costs and trying to figure all that out. Looking from the outside in at the industry, it seems like everyone’s crushing it with the cloud, but the reality is most enterprises have a lot of work to do to get really proficient with any cloud provider, whether it’s Azure, AWS, or Google.
Patrick: That’s a really interesting point, and I think we all heard a lot about that at Microsoft Ignite this year, where people were asking really technical questions. I’ve had a couple of folks say that they were skipping sessions and instead going to that giant Microsoft Meet the Engineers area and just having the conversations that you would normally have after a presentation right there, because they’re in the throes of real implementation.
Mike: Yeah. We’re seeing more of that, too. It’s really interesting because for many years, the trainings that I was doing were prerecorded. Lots of pre-production and very little interaction with students. Now it’s flipped. Most of the stuff we do is live and interactive in Q&A, because there’s so many questions. It’s very difficult to answer the questions when you’re spending so much time pre-producing content. There’s still many use cases for that, but to your point, it’s like, there’s so many questions in the marketplace that you can’t wait long enough to pre-produce content. You have to get in there and just start helping people.
Mike: So, the opportunity has been awesome and I think that it’s going to be like this for a while.
Patrick: Yeah. That was one of the things that happened in the SolarWinds user community, THWACK, was that there were a lot of conversations, to your point, about that long tail of really specific questions around it. So, that kicked off a survey, and it turns out that over 50% of the users on the community are actually on Azure, not on AWS. So, enterprises seem to be waiting a little bit more for tooling and process.
Mike: Yeah, absolutely. This is one of the things that I’ve been talking about a lot lately. If anybody has heard me on another show, you’ve probably heard me say this, but when I started in IT, it was the ’90s, and I had to make a decision in that early timeframe of, am I going to go be a Novell expert, a Novell NetWare expert, or am I going to be a NT 4.0 expert? All of my homies that were a couple of years ahead of me were like, “Come to NetWare. Get your CNE certification, your certification”–
Patrick: It’s awesome on token ring.
Mike: That’s right. Yeah. I just had a gut read on it. I’m just like, “I don’t know, man. I think that the Windows stuff is cool.” So I went down the NT 4.0 path, and then Microsoft came out with Windows 2000 and Active Directory, and they destroyed Novell.
Mike: So, I think that if you look at that history, Microsoft is… They’re so good at building management interfaces and they’re so good at policy, and so from where I’m sitting now, I think my prediction will be Microsoft will dominate in the enterprise because they have that background. They’re already starting to do a lot with governance. In my opinion, they’re starting to lead in that area, especially if you look at things like Azure Policy. So–
Patrick: Mhmm and Blueprints.
Mike: Yeah. And Blueprints, and the whole RBAC story, and there’s, like… for some things like that in other clouds and stuff, but it’s Microsoft’s proficiency, and since they’re already penetrating the enterprise so well, and there’s tons of Active Directory out there, to me it’s obvious that’s where it’s going. Happy to eat crow if I’m wrong, but I don’t see it happening.
Patrick: Well, okay. So, that brings me to another question, which is—and you certainly have the background for this— would you say that all cloud providers are exactly the same? They’re all created equal? Or is it increasingly a little more difficult to really assess how they’re different?
Mike: Yeah. I think they’re very similar and, terminology aside, if you really get into it, they’re all very similar. If you look at AWS and Azure, there’s almost complete parity at any given time. Not complete, but nearly, right? Like-for-like services that all do similar things and just have different names. So I won’t say that all cloud providers are the same, but they’re all offering similar services and they all have their own strengths and weaknesses.
Mike: Like Google, their strength is big data, they’ve been indexing data for 20 years. They’re experts at that. Right? So, they’re probably going to do well in the big data space, compared to the others. AWS has been amazing because they’re first, they’re good with startups and non-Microsoft shops, but there’s a lot of Microsoft on there.
Mike: But again, I think Microsoft’s going to dominate in the enterprise because of their management capabilities and the policy stuff. But at the end of the day, I think if you get good with one, it’s going to be an easier lateral move to another. It’s just like learning PHP and then going and learning JavaScript after that. Right? It’s almost like–
Patrick: I’ll take JavaScript over PHP.
Mike: Yeah. But there’s similarities, right? And so it’s like, once you know one really well, then it’s like, “Okay. I can pick this up over here because I understand looping and if statements and all this stuff.” So, I think if you get good with one, it’s going to be easier to make a lateral move.
Mike: I think what will happen eventually, over the next several years, is, they will become more commoditized, and the management layer will be where it’s at. It’s like, people talk about multi-cloud, and so we’re going to see a lot of multi-cloud products come out on the market, and I think whoever can bridge that gap across all the clouds is going to be in a really good spot.
Mike: Right now, they’re not all identical. They’re similar, but that’s my feeling on it.
Patrick: Yeah. It was interesting watching the keynote a few days ago for re:Invent, where, in terms of whether you’re an early or late adopter with cloud, I think AWS has up to four now, container, either delivery or orchestration products. Is that kind of differentiation for them maybe a product of those early adopters who have a particular investment in a cloud-native service, that coming a little bit later to it means you’re more likely to get something that is a second-generation implementation, or is designed to be easier to use, versus easy to get started with?
Mike: I think they’re all doing a better job of making it easier to get started with. When it comes to AWS, one of the things that they did a little bit differently when they started their container support was they built their own orchestration system. So they have ECS, right? The Elastic Container Service where you can just go and run your Docker containers, and they got criticized for that, in the early days, for doing that. Then they created a flavor of hosted Kubernetes, or managed Kubernetes, called EKS.
Mike: So, it depends. I think the managed service route is going to be where everything goes. They’re all going to try to lower the barrier to entry and make it as easy as possible. So, building your own clusters of servers on virtual machines and having to deal with availability and backups is going to become less and less of a concern. They’re going to slowly manage people into more managed services. That’s just where I see it going, so I think it’s going to get easier going down the road.
Patrick: What would you say the biggest problems are for users when they’re migrating applications beyond just lift and shift, to really take advantage of the services of the cloud?
Mike: Yeah. In terms of not just lifting and shifting, but transforming into something else, something brand-new, like microservices and containers and all that, I think a lot of companies need to be careful. There’s been several, right now, that have put an enormous amount of resources into things like trying to split up applications into microservices and trying to do Kubernetes and things like that, which is great if you have the operational maturity and your engineering team to do that.
Mike: But a lot of our customers are mid-size, 500 users, 1,000 users, 2,000 users. Generally, in that type of scenario, the engineering teams aren’t as good as the engineering team at Netflix, so it’s like–
Patrick: Yeah, Netflix has more employees than some of these businesses do.
Mike: Exactly. I think that people need to be careful with that and really understand, do we have the ability to do this right now? And if you do, great. Knock yourself out. See if you can add value by transforming the application, but also not worry about feeling obligated to do it.
Mike: I think the second thing, too, that I would tack onto that is, a lot of the companies that we’ve worked with… well, not a lot, but several have… it’s been obvious to us, as we got started in projects, they didn’t really have a clear goal. They just feel obligated to get in the game. And that’s great, but I think what companies should focus on is what’s painful and something that’s easy to do first, and then get some points on the board, and then get into transformation of applications to cloud-native and all that kind of stuff. So, that’s kind of–
Patrick: Decent visibility, low enough risk, small enough in scope.
Mike: Yeah, 100%. Start off with offsite backups or something basic, just to get the feel for it so you can get your arms around the cost controls, the security, and then crank up the intensity from there. But you’ve got to be honest with yourself, look at your team, and realize, what can we actually accomplish? Because if you weren’t doing microservices and containers before, it’s going to be very difficult to pull that off in a quick turnaround timeframe.
Patrick: Well, I think you also hit a bunch of the points that go to people feeling a little bit nervous about it, and it can’t be all gloom and doom and impossible, because the reality is most companies, when they start, pretty quickly accelerate, and then you’ll start hearing them say in a believable way that they will be all or majority cloud within a reasonable amount of time. So, obviously they are having some success. So, do you have any stories, maybe, where a company was surprised at how successful they could be with the migration?
Mike: Yeah. We work with a software company that they have to spin up QA environments to test their code, right? Well, they don’t have to, but what they were doing in the past, they had a QA environment that was just sitting there, and every time there’s a new version of their application, they would push it to that environment, and just having that environment sitting there over time, we all know how it is, right? The server gets beat up over months or years. It’s just not the same as it was the day that you installed it.
Mike: So, what we did for them is we created a CI/CD pipeline process for them to where, when their code changes come in, it spins up the staging environment brand new, right? So, it’s following that immutable infrastructure pattern where we bring up the environment and we test our code against it and we delete it. When it’s in production, the idea is in that environment, if something messed up, we could just replace it instead of babysitting servers and logging and doing that kind of stuff.
Mike: For them, it was a game-changer because every time they test their application, it’s tested against a pristine environment, so we know that there’s nothing weird lagging over the last couple of months or whatever. I think that they might have underestimated the value in the beginning.
Mike: A lot of what, I think, we fight in working with our customers as an issue, if you dig down into it, usually it’s not the technology. A lot of times it’s just the human side of not wanting to change, and that’s the business that we’re in.
Patrick: Yeah. When you’re in the business of high risk of outage, where you’re told you absolutely have to maintain as much availability and resiliency as possible, but at the same time, you are delivering commodity services in a cost center, that’s never a mix for feeling eager to take a lot of risks.
Patrick: Here’s a question for you. That aspect of delivering applications, of mitigating risk, especially when you get into those unknown unknowns… I don’t know. Rumsfeld quote. That when you start, to your point of transitioning application, or even just maybe modernizing an app, where you’re going to take a layer of that application and convert it into microservices, managing the delivery successfully for that application changes a lot, right? It’s not lift and shift. These aren’t the same packaged applications that people have been using for a long time, with well-established, decade-old management interfaces you can just connect to.
Patrick: How are teams adapting to that, to where something that used to be easy to get access to is now opaque, or they’ve added cloud-native technology that maybe has little or no robust management interfaces for it, where ops is almost becoming developers, or at least taking a more dev approach out of necessity… How are they getting over that?
Mike: Yeah. It’s not easy. It’s a big hill to climb, if you haven’t done that before. I think even if you’re, from a operational perspective, good with containers and supporting cloud-native applications, there’s still the challenge of how do we get visibility into what all these services are doing? How do we trace issues, and how do we see what’s going on?
Mike: Honestly, it’s a big problem right now. There’s products on the market that aim to solve this problem, and many of them are super effective. People are using them. But I think it’s an education challenge, right? People have to have, coming up front… If you’re in operations and now you’re supporting microservices, if you don’t understand the basics of how HTTP works, then you’re not going to understand the basics of how are these services interacting together. That’s a core piece, before you even get into installing, monitoring, tooling, and things like that, in my opinion, because you wouldn’t even know, necessarily, what you’re looking at.
Mike: I think one of the bigger challenges is that, is how do we monitor all this stuff when it’s so distributed? It’s an interesting time, and I think that the industry still has to close the gap on that a little bit.
Patrick: Walk us through an example of what that looks like. If you looked at a regular three-tier application, right? You’ve got a database, a data layer, you’ve got some business logic behind that. You’ve got a service layer, and then a firewall, something that would run on-prem. Right? Something that we’ve all delivered for a really long time. How does that application look once you migrate it to cloud with some modernization or transformation? What do those typical, first hello production look like when they’ve migrated?
Mike: Yeah. I think one of the things that people need to do is try to lead with the managed services because that makes it easier. But if you’re taking a typical three-tier application, you can mimic that architecture by going strict… infrastructure as a service of doing virtual machines, and all that kind of stuff. That’s very difficult to pull off.
Mike: To answer your question, it depends on the approach that the company takes. I would like to see a lot more customers try to go down the managed services route, because it reduces a lot of the work, but a lot of companies don’t realize the amount of additional effort that takes place when you’re supporting large architecture like that, and it’s all virtual machines. There is a lot of little things that connect together. There’s a lot that can go wrong, because now you’re responsible for high availability and all that kind of stuff, and it’s not easy to do.
Mike: The idea with this would be to make it as easy as possible. Go managed service if you can. But sometimes it’s a blocker. Sometimes you need root access to the servers and you don’t have the time to rebuild the application or the way that it works, and so you’ve got to rely on the old-school way of doing things. We see that a lot. We see a lot of people moving virtual machines into the cloud. But then, in that model, it’s like, well, did you migrate to the cloud, or did you just move a server from one data center to the other?
Mike: I think companies need to do a little bit of a better job of focusing on, what’s the value add in the cloud that we can extract before we just take an architecture, just move it up there? Right?
Mike: The way it would look depends on the cloud and what technologies they can include, but basically, the idea is, especially if you want to isolate your applications and stuff, and this comes in the conversation regardless of whether you’re doing managed or full IaaS, is putting stuff in your own virtual network, just like you would on any virtualization platform, isolating it, and then deploying your resources into that. There’s a lot of moving parts, depending on the platform that you’re talking about.
Mike: You might be able to mix and match some infrastructure as a service resources with some managed services. Maybe you’re still running your web applications on VMs, but now your database is a managed service, and you don’t have to worry about backing that up and patching it because the vendor’s doing it for you.
Mike: It really depends on the customer, but those are just a couple of thoughts there, shooting from the hip.
Patrick: There’s a lot of confusion around multi-cloud and whether that basically just means you’re hybrid and you have either multiple instances of the same subscription, or you’re actually using different providers, and then some people think of that as an agnostic set of platforms with portable services. But is multi-cloud here to stay, in whatever form it is, or do you see customers actually developing gravity around a particular vendor and sticking with it?
Mike: Yeah. I think a lot of people that talk about multi-cloud are talking about it in the sense that, “Hey, we want to eliminate the vulnerability of going all-in on Azure, and now we’re stuck there.” I think, to be honest, 2020 is not going to be the multi-cloud year for most companies, because most companies aren’t even good with one cloud yet, and so it’s like, let’s get one under the belt and then talk about the other one.
Mike: The second thing is the management story, it just isn’t there. If you’re doing multiple clouds, you’re literally on your own, connecting the stuff together. If you’re trying to span an architecture for an application across multiple clouds, you could pull it off, but you’re going to be all completely on your own, and there’s not going to be, right now, a management layer, necessarily, that covers the entire board.
Mike: There’s companies like HashiCorp that are starting to build really cool things that span all these different platforms. Their Infrastructure as Code solution, Terraform, can talk to all the cloud providers. They’ve got service mesh. They’ve got all kinds of cool stuff, and I think that we’ll see them and others come into the market and add some multi-cloud management support, but it’s super early, and I don’t think that anybody could really pull it off effectively anytime soon. That’s just my opinion.
Patrick: Is a big part of that… Especially folks that I speak to, where they discover that they had a whole bunch of shadow IT, they had lots of different accounts, and they’re trying to make sense out of that, but it sounds like there’s just a lot of experimentation that’s going to happen while production has to continue. Is that a big part of what’s driving application performance management as a tool, versus what would have been a lot more infrastructure maybe bolted directly to a single provider?
Mike: Yeah. What’s happening is that all the providers are in this model, now, where they’re competing on how many changes they can make every year. Right? So, if you go to the annual conference for any of these big providers, they’ll tell you how many things they changed that year, how many services they introduced, how many features were in there.
Mike: We’re literally in the era of continuous delivery, right? These technology companies are building that stuff that way. That’s why it’s impossible, almost, to keep up with what’s happening every single week, and things are changing in the UI almost every week.
Mike: What’s going to happen is, customers are going to have to level up, right? Because the technology companies are building their software that way. The solutions that they’re building to give to their customers fall into that same paradigm. Customers are going to have to start to get into that iterative process of not just standing stuff up and babysitting it and playing defense for three years and then do another project. It’s a continuous thing.
Mike: So, experimentation… Companies are going to have to adopt a learning culture. We talk a lot about this and in a world of DevOps, right? Making it safe to fail and having a learning culture and giving people time to ramp up. All that stuff. If these technology companies are going to pull it off, they’re going to have to put pressure on their customers to play up to their level. It’ll be interesting to see, over the next couple of years, how Microsoft, after their own culture transformation, gets their customers to do that. Same thing with AWS and Google.
Mike: But, yeah. That’s the game that we’re in now. It’s try something, measure the feedback, and then let’s try and improve it again. It’s every day for the rest of your career.
Patrick: Right. So, you’re gonna monitor, like an admin, but really you’re going to start to think more like a user and whether they’re actually really delighted by that application, not just meeting SLA.
Mike: I think you’re going to monitor like an administrator, and then you’re going to iterate like a developer, regardless of what you’re doing.
Patrick: All right. It almost sounds like the winner among clouds, or at least those that will be real successful, are the ones that are not just providing technology, but really helping to change that traditional mindset toward a learning culture, toward a faster, a smaller delivery. When we talk about DevOps, we talk about agile, that these are human changes or cultural changes, and certainly enterprise, and IT in particular, has a lot at stake there.
Patrick: So, what’s a great way to help shift that mindset? Because you can’t do it by dictating policies in a big, broad team meeting. It seems to be a little bit more one-on-one. But how are you seeing that work on the ground?
Mike: Well, we’re seeing a lot of challenge, to be honest. We’re not seeing it work very good, to be honest with you. But what’s going to have to ultimately happen is the leadership of the company is going to have to come in and say, “Hey, this is the direction that we’re going.” The whole thing with the finger pointing and the blame-game stuff and not working together just slows you down. We’ve just seen this so much with our customers.
Mike: These cloud platforms are not going to be as profitable as they possibly could be if their customers can’t move as fast as, or at least closer to as fast as, they are. The only way to move faster is to increase collaboration and to help people learn faster. That’s going to have to come from senior leadership of companies.
Mike: If you’re listening to this and you’re an executive, that’s what it’s going to take. If you want to be successful in cloud and not spend too much money and have happy people that don’t quit, you’re going to have to deliver that, and the people that succeed are going to be good at that, and the ones that pay too much and aren’t good at it are going to have a bad culture.
Patrick: Well, exactly. I think Gartner actually had said that something like 75% of enterprises were going to experience actual visible outages as a result of things like skills gaps. Not even what they don’t know or maybe experiences that they don’t have, but even skills gaps.
Patrick: If you were you, rewinding back to today, what skills would you start with in terms of recommending our listeners focus on?
Mike: I think probably the best thing to focus on right now is core services. If you are good with compute, storage, networking, and security, every cloud service is predicated on those four things. Getting started in… if you look at Azure and even AWS, they have really great entry-level education and certification content. I would get started there and then really focus on, how do I become proficient with the four pillar services that everything’s predicated on? How am I going to get good with compute? Which is beyond just virtual machines, but also things like serverless functions and things like that. The managed services that are powering some of the Kubernetes services could sometimes be considered compute.
Mike: But understanding compute, understanding storage, understanding networking, and then understanding the identity and access management piece, the security side, if you’re proficient in those areas, then you can move over laterally to something like a managed SQL implementation and understand what’s going on behind the scenes, and understand how it’s secured, and how it integrates with the network.
Mike: That’s what I tell all my folks, and when I bring somebody into our team, that’s one of the first things that they’re going to do. They’re coming from an AWS background, now they’re going to do some Azure stuff. It’s like, let’s get good with these four services in Azure, and then we’ll crank up the intensity from there.
Patrick: But it sounds like those four services… and it would make sense why they’re successful… are aligned with skills that a lot of people already have. So, really, you’re saying that it’s basically differences training as much as anything else?
Mike: Yeah, it’s like, take you know now, and how does that translate into the cloud? Then, once you understand how they’re doing it in the cloud, then you can start to go into some of these more nuanced type of services that might seem foreign.
Mike: But that’s a good point. A lot of what you know now as an infrastructure person, if you are that, you already know a lot about the cloud. You just have to figure out what they’re calling it, how they’re implementing it. It’s easy to get intimidated by the breadth of services and all those different things, but the reality is, most people won’t tell you this, but I’ll tell you this, as a quote-unquote cloud expert, I don’t know everything and I never will, and I’m super comfortable with… and I love it.
Mike: If you can sit there and be like, “I love that new services come out every day, and I’m constantly confused,” then you’re going to really enjoy working in the cloud, because nobody else is ever going to know all of it. Even the people that work at these places that I talked to, they know their thing, and that’s it.
Mike: So, yeah. That’s what I would say.
Patrick: Okay. So, we’re getting ready to wrap up, and I just wanted to bring this back to tools that are helpful in particular application performance management. Why don’t we use that as an example of taking skills that you already have? If you were used to looking at, let’s say, page delivery time or components, and managing what’s coming from a CDN versus what’s coming out of the JSON services API versus the front end, for example, in an application where you would typically rely on infrastructure monitoring behind the scenes, how do you adapt that that type of APM for cloud?
Mike: Yeah. I think you have to be tuned in to what it is, which cloud platform you’re using. So let’s take Azure for an example, since, based on your survey, most of the folks that you guys are talking to are Microsoft-focused. When you look at Azure, the platform itself has a native monitoring service. So Azure Monitor, which is really tying in lots of different capabilities, but at the basic level, for traditional infrastructure, you’re going to have the core metrics that you would always expect to see, things like CPU utilization and memory and disc performance, all that kind of stuff, and in some different ways depending on the type of service you’re looking at.
Mike: So, there’s that built-in stuff. You can also customize it. Then from an application perspective, if you’re building web apps and so on, at the application layer in the Microsoft world they have something called Application Insights, which is an insanely cool telemetry application performance monitoring service, where you can get lots of insights into everything that’s happening with the application, requests that are coming in, and is your service throwing errors, all that kind of stuff.
Mike: The one thing I will say, though, is those services have been a moving target, right? The different capabilities are coming in. The core stuff is there, like the basics of how do we monitor the core infrastructure, the stuff that we’ve always monitored? Like, AWS has CloudWatch and Azure has Azure Monitor. But when you start getting into more of these newer type of services and newer ways of doing things, that’s going to require some updates to the way that we’ve done things in the past.
Mike: So, if you haven’t supported a web application in the cloud, you might get sucked into something like Application Insights, or you might get pulled into something like AWS X-Ray, to kind of get in the depths of application performance monitoring there. It really depends on the platform that you’re on, the application that you’re supporting. If you’re doing traditional VMs and stuff, there’s great support in the monitoring systems, and even if you’re doing custom applications, but you have to know what platform you’re on and what tools and services are available in that platform, because it’s going to be very different from one to the other.
Patrick: So, in terms of application delivery, and especially application performance management, it sounds like you’re saying that, as a set of practitioners, maybe we are a little behind in terms of our expectations for parity with capability of tools, but that vendors are making a lot of strides toward best-of-breed solutions, integrated solutions, and ones that are a lot easier to get started with for managing applications.
Mike: Yeah, definitely. I think that, going back to what I said earlier, as time goes on, the cloud platforms are going to continue to become more and more commoditized, and the people building services on top of something that can work across all of them are going to win. So whether that’s managements, whether that’s performance monitoring, whether it’s some kind of single pane of glass or some other thing that we’re not even thinking of yet, I think those companies providing those tools and services have a huge opportunity, because right now, if you’re looking at a lot of this stuff, especially if you look at Azure, right? Just taking them because we’ve been talking about them a lot, right? They have their own native monitoring solutions, but if you really do start dabbling in multi-cloud, then you’re going to need something that can talk to the other clouds. Right?
Mike: There’s a huge opportunity there, and I think that it’s also going to give people maybe even more depth into the things that they’re monitoring, the different management strategies that they’re putting into place. So, yeah, it’s going to be interesting to watch things unfold.
Patrick: Well, awesome. Mike, as we get ready to wrap up here, is there any last piece of advice that you might want to share with the audience about assuring that the applications that you’re delivering well right now on-prem are also going to make people happy once you’ve migrated them over to cloud?
Mike: Yeah. I think my advice might seem weird for folks, but I’m going to throw it out there anyways. My advice would be to start looking for ways to become a better communicator and become more conscious of what the people you’re working with are actually seeing, because going back to some of the things we’ve talked about, the biggest, the absolute number one thing that I struggle with when I work with companies is the human side. If we can get to a place where you can actually consider someone else’s point of view before you make a snap judgment and those types of things, it’s going to be a lot better off. So, that’s my biggest thing.
Mike: Then also, as a practitioner, as things continue to evolve, the skills are going to continue to be more and more commoditized as well. At some point, the way for you to really add value and stand out is not going to be because you’re such a great technologist, it’s going to because you’re such a great communicator.
Patrick: Well, that’s great, and that’s also all the time that we’ve got today. If you’d like to learn more about APM tips and tricks or how to be successful managing your hybrid multi-cloud environments, be sure to check out the Orange Matter blog. It’s full of expert insights and guidance. Thanks for joining us on SolarWinds TechPod. Mike, it was great to have you.
Mike: Thanks a lot. It was really great to be here. I appreciate it.
