PurePerformance - Perform2020 Andi on the Street: Observability & Beyond, AI powered Kubernetes, and Real User Monitoring

Episode Date: February 6, 2020

Andi Grabner, our man-on-the-street, gets the scoop on:-Observability and beyond - with Thomas Rothschädl-Automated, AI-powered answers for Kubernetes with Matt Reider-RUM Roadmap with Alexander Sommer

Coming to you from Dynatrace Perform in Las Vegas, it's Pure Performance! Hi everybody and welcome to Pure Performance and PerfBytes, coming to you from Perform 2020 in Las Vegas. Our man on the street, Andy Gravner, has sent us a few interviews, which we've compiled in this episode. There's a short musical interlude between each, so be sure to stay tuned in. Take it away, Andy. Welcome, everyone, to another episode of Pure Performance Cafe in Vegas at Perform 2020 at the Cosmopolitan. And I just bumped into another colleague of mine. Thomas, hi! Hello Andi, hi, how are you? Very good, very good. It's been a busy week here. Definitely, definitely, yeah. So Thomas, can you quickly give me an introduction about yourself

and also your role and kind of like how long you've been with Dynatrace so that people know who I'm actually talking to? Yeah, my name is Thomas Rothschildl. I'm with Dynatrace so that people know who I'm actually talking to. Yeah, my name is Thomas Rothschild. I'm with Dynatrace now since five and a half years. I'm a product manager for quite a lot around the service area. So I'm in charge for Node.js, for the IBM integration bus, and also for AWS Lambda. We are now heading forward here in this area. Awesome, perfect. Hey, Thomas, you have a session here at Perform

around observability. Now observability is a big topic. I mean on the one side it seems like people, especially some of our let's say you know friendly competitors are selling it like it's the latest invention that came out of Silicon Valley but But obviously, we've been doing this stuff for a long, long time, thinking back about the PurePath and how we catch the data. Exactly, we invented a big part of that to give our customers the observability into their systems. But I mean, it's great that actually it has a lot of awareness now. So can you tell me a little bit about what you're going to cover in your breakout?

Because remember, not everybody will be able to see it live. A lot of people will hopefully watch the recording. So what are you talking about? What are the key takeaways? Exactly. So the first point, we start grabbing the current buzzword, as you mentioned, observability a little bit to draw what you can achieve with it, how it works. And then I led over to what we do out of the box with Dynatrace One Agent,

how our customers can use everything that we collect out of the box with Dynatrace One Agent, how our customers can use everything that we collect out of the box. And the goal is also to show, yeah, it's not just to make some specific points observable, to analyze log files, to see some metrics. It's the most important things to have this in one product, to see this together, to really do good conclusions, why is the service low, why is the application not performing, and how this can be achieved in Dynatrace, on the one hand, automatically, but also how this data can be enriched by adding specific points into their own environment. Perfect. So that means you're going to talk about how to instrument, let's say,

applications that maybe where the one agent doesn't fit. Exactly. So the one point, how this visibility can be manually achieved and can be combined with the data the one agent how this visibility can be manually achieved and can be combined with the data that one agent automatically gets out. I also want to recapitalize there a little bit to get an understanding what we do automatically. Because I think this is very important in this microservice environment. There are lots of new startups there. They explain the inventive world on the one hand side.

But when you do it in that way it's quite a lot of manual effort, so I also want to show this to give our customers a good overview how, if this is worth the effort, and also the maintenance to have it there and what do they get out of the box with Dynatrace, like how the method hotspot is working out of the box, how thread analysis can work out of the box with Dynatrace already without any change there. So to compare this a little bit to get a better feeling, where is it worth the effort on the one hand side?

On the other hand side, what do you already get with Dynatrace, although it does not have the observability stamp on it? Yeah, yeah, yeah. So, I mean, there's a lot of, as I said, it's a big topic. We have it in multiple breakout sessions. I think there's Daniel and Sonja talking about OpenTelemetry, which obviously feeds into

the topic as well. Exactly. It's one part to give observability by tracing to applications, but it's just a very small part of this. Yeah. Then you just mentioned microservice environments. I think Alice Meyer has been doing a session on observability, especially for microservice environments.

That's great. Exactly. In the beginning, you said product manager, and I think you mentioned service bus. So also these systems where maybe manual instrumentation is not possible, and the monitoring is not possible. I would assume you will also talk about plugins. Exactly.

So plugins will come, will be covered there. And also one point is how log files can be aggregated there and collected. Perfect. So obviously a hot topic, everybody should watch your recording. Anything else that you want to kind of explain here, any other things why people have to watch? I mean it's a lot already, but if anything comes to mind. Yeah, I'm really also glad not just to present there in the end of my session, I also want to have a quick discussion with the audience there.

Perhaps we can, yeah, the five minutes there, to also get some feedback there from the audience. That would be really great because this also enables us to improve our product in the future. And that's also what it's performed about, to get into close contact with our customers, with our upcoming customers, also to see how they are using observability.

So this would be also great. If someone is working there, join the session and get in discussion yeah and then in case they cannot make the session because there's somewhere tied up i'm pretty sure you will also be at the innovation lab like the towers right in the expo area exactly that's a great and you bring up a good point there's a lot of product managers here from dynatrace so let's make sure that everybody leverages the opportunity to speak to you directly and kind of bring forward their individual requirements and situations. Exactly.

That's very important and I'm really looking forward to this. Awesome. Hey, thank you so much. Enjoy the rest of the show. Thank you very much. Okay, see you later. Bye. Hey, welcome everybody back at Perform 2020 in Vegas and I'm running here through the

hallways. I found a place next to a coffee machine as you can hear from the background noise. But I also bumped into Matt. Hey Matt, how are you doing? Good Andy, how are you? Good. Hey Matt, you are I think kind of new at Dynatrace. I mean, kind of, quote-unquote.

Yeah, yeah. A lot of people start. When did you start with Dynatrace? I started with Dynatrace, I think, in August. So I moved from Silicon Valley to Linz. You moved from the original Silicon Valley to the Silicon Valley of Austria. That's right, yeah. From what I hear, that's going to be more and more common as time goes by.

But yeah, I'm blazing the trail for other folks to leave California and go to the tech hub of Linz, Austria. Here we go. Perfect. Well, it's great to have you on board as a product manager. You are responsible for what areas? Yeah, so my background is mostly in PaaS, so Platform as a Service.

So here with Dynatrace, I'm focusing on that. Cloud Foundry, OpenShift, and Kubernetes. Perfect. That also then makes sense now to me why one of your breakouts that you're doing later today is called Automated AI-Powered Answers for Kubernetes and Container Monitoring at WebScale. Yeah. A lengthy title, but I think it hits all the major points

that we want to bring across. WebScale. Yeah. A lengthy title, but I think it hits all the major points that we want to bring across. That's right. Yeah. So if people, not everybody has a chance to actually watch the breakouts live, they are fortunately recorded. So what will people hear and see in your breakout? Yeah. So we're really fortunate because we have Jeff Getzloff from Walmart who will be doing

the presentation with me and so you'll be hearing and seeing some real-world examples of how a company is using a company we all know yeah very large enterprise using platform as a service so kubernetes on a huge scale so actually they're deploying it to all their brick and mortar stores, and they're powering their POS systems within those stores. So they run Kubernetes clusters in their stores? They do. So not only do they use Kubernetes for all sorts of things within Walmart, within IT,

and their public websites, and all sorts of different departments, but they're actually deployed at the stores themselves, which was news to me. But Jeff has talked to us a little bit about how the challenges that that is presenting them and the problems that Dynatrace can help them solve. And specifically in the context of Kubernetes, he's going to show everyone in the audience how he uses Kubernetes to identify problems and to solve those problems. And he's also going to preview some of the new features

that we're excited to launch that we're actually announcing this week at Perform. That's awesome. I actually remember I did a performance clinic with Alois Meyer, who I had on an earlier podcast. We did one, I think it was back in probably November or December, when he talked about the latest things that are coming out of Dynatrace when it comes to Kubernetes monitoring.

So could you kind of maybe glance over quickly what are the new things that people can expect? Because I know a lot of our customers are moving to Kubernetes. So what's new? So one of the biggest sources of concern for anybody who's running a cluster, everybody talks about the cloud and thinks that there's no capacity planning when you're thinking about the cloud. But that's, of course, when you're in someone like Jeff's shoes, capacity is really important. And so one of the new views that you'll see within Dynatrace's dashboard is you'll be able to see the utilization of a

Kubernetes cluster and see basically how it's performing as a whole. Rather than looking at the workloads or the containers or the nodes, you're actually looking across the entire cluster to see basically whether you need to add resources to make the cluster work a little bit more smoothly. So that's one thing we're adding. And that has all of the requisite features that Dynatrace offers. So things like management zones are there,

and you can also sort of look at a timeline and zoom in and the things that you would expect that are available with other areas of the product. We're also showing some new features with events. So Kubernetes Invents ingest. You'll be able to, for lack of a better word, subscribe to events. All of the events that if you were to go to Kubernetes, the kubectl command line, and you were to basically, there's

a command that allows you to see every single event. There's hundreds of them. And you can basically pick and choose which of those events you wanna bring into Dynatrace. And that's really helpful because then you can correlate those events with problems, which is really why you'd wanna see those events, is you wanna see, well, things are spiking, or I saw there was

a response problem here, and that's probably due to this event that occurred. And maybe that event had to do with an app wasn't starting. And so the node was flailing around on a certain, or a pod was flailing around on a certain node. You'll see those events because they'll all be ingested. And I would assume, obviously, as everything what we do in Dynatrace flows into Davis.

And it basically enriches the smartscape. That means we see all these events. but the most important thing is Davis also sees these events and in case there is a problem it automatically picks it probably up. It's a root cause. Yeah so this is as you said I'm new to Dynatrace but this was one of the things that really excited me about Dynatrace. It's really not just data. What you're doing is sure you're ingesting events and logs and metrics, but really the power of this is that when a problem occurs, Davis is understanding what these relationships

are to show you exactly where the problem occurred. That's pretty cool. So if I can recap this, first of all, we have a great speaker from Walmart. He's actually going to show us how at large scale they are running Kubernetes, why they run Kubernetes, but most importantly how they monitor with Dynatrace and which problems it actually solves. And then people will get an outlook into what's coming from them soon, which means better support for infrastructure monitoring on the Kubernetes side for better, let's say, capacity planning, all these things you just laid out.

Events are very important so that actually Davis has more data about what potentially caused a certain problem. That's right. I mean, that's awesome. It is. I guess the only thing that I'll say that we're neglecting to talk about is that we also have been hearing from our users that while they care about containers and they care about pods and workloads, which is a term that's used just in Kubernetes, really

what they care about is the application. And so we have a new view, which is a view that we're basically devoting towards cloud applications. And it's a new part of our entity model. So you don't care about some obscure set of containers that have some weird names. You want to bind all of these things together in groups, and they're named logically according to the things that you're deploying, which makes a lot more sense. So you're going to be able to see a list of all of your applications

and some of the settings that you've put in those pods, like how much CPU you devoted, how many instances you wanted to have versus how many are actually deployed. A lot of the same types of concerns that application developers have when they're on this new platform called Kubernetes. We will give you a nice entry point into looking at your applications and then drilling down into that to see how they're performing. And now you would get me extremely excited if you also would tell me that these new cloud

applications also support things like blue greens, canaries, and things like that? Yeah, so this is where we're marrying applications with DevOps, right? And that's what Kubernetes is good at. So when you're deploying an application, you could do these in stages, like a canary deployment. And you could see within two different sets of applications,

you know, maybe one you're seeing if the health of that application is performing so that you can then continue to deploy more instances of applications. Maybe one you're seeing if the health of that application is performing so that you can then continue to deploy more instances of it since you know that it's less of a risk. For sure, blue-green deployments, and that all ties into a topic that you know so well, which is captain. Exactly. This is our road towards moving from AI

to actually potentially automating the solution for a problem or deciding some logic that says, okay, this isn't performing very well. Maybe we want to roll it back. That's sort of the next phase, logical phase of how observability might move. That's awesome. Yeah. Hey, Matt, thank you so much. I know you have to run because your session is starting soon.

So all the best for the session session first of all. Thank you. Great to see a guy from Silicon Valley moving to Linz. Hopefully more will follow as you said. It's hopefully it's a good place to live for you. We've moved with the family, right? Yes, my whole family is in Austria and we're taking advantage of it. We're going skiing. There you go. And we really love it. It's been a really good experience so far. Awesome. And now, after hearing the barista really working hard in the background, I think we should get a coffee.

Sounds good. Let's do it. All right. See you later. Bye-bye. Thanks, Andy. Ciao. Welcome, everyone, to another episode of Pure Performance Café, still in Vegas, believe

it or not, it's a long conference here at Perform and the Cosmopolitan. And I just bumped into another colleague of mine, Alex Sommer. Hi Alex, how are you doing? Yeah, fine. How are you? Good, good. I think we're both, it's a challenging week, right? Long days, but good conversations hopefully.

Yes, it's always at Perform, so many customers, so many feedback. Too much for us, but it's really cool. Yeah, but for you, I mean, feedback is obviously very important because, Alex, for those people that don't know you, maybe you want to quickly explain your role within Dynatrace. Yes, so I am responsible for the digital experience management roadmap, working with the PM teams on the roadmap and of course trying to bring new features to the product so that you can

solve your use cases. That's awesome and you've been with the company for many, many years as I remember. Yes, I think now it's the tenth year so I'm not sure but I have attended a lot of performance and every year is getting better. That's cool. And so you said digital experience management. So that means everything that people know around RAM, real user monitoring, whether it is your digital properties from like your browser, mobile or desktop. I mean, there's so many things around digital and a lot of things is happening in the digital

world. And I think your session is a roadmap session which is as far as I know one of the biggest or the heaviest attended sessions overall so that's why I believe a lot of people will be very interested in seeing and hearing what you have to say so what is it that you are going to show and tell yes yes Andy you are right so I'm talking about the roadmap on digital experience management so as you stated out this is mainly about real user monitoring, but also, of course, synthetic and some user behavior and, of course, our session replay.

So what I will do is, as I have done in the last sessions, I will look back a little bit what have we achieved in 2019 and what will come in the next months to the product and of course as always you looking at real user monitoring for web applications for mobile applications but also for synthetic on the synthetic side and of course for session replay that's very cool are there any I mean we have a lot of customers that have used RAM for many many years probably back in the days when it was still appmon and it was real user monitoring and now obviously in Vantage

with the real user monitoring what are some of the new things that people may have not yet seen and what are the new things that you are going to talk about yeah so as always we we have when we go to perform we have our innovation tower where we show new things which are already in development or even in the product but not rolled out. So we will show their new things like increased number of session properties with some out-of-the-box properties to really use the segmentation on the user sessions for some standards which we can achieve over our new features. But what we also will do is we will show you the big things we have done in 2019. So because I think we have released over 50 blog posts only in my area, so I'm not sure if everyone knows them. So we will give you the demos on really the new features,

but also give a little bit an out-blink what is coming. So for example, the focus we set. So what we learned over the last two years is that perhaps we need to rethink our JavaScript code which runs into the web applications. Because in the last months what we have seen is that more and more of our customers are going to very modern web applications and we do not have the out-of-the-box visibility what we want to do. So this is perhaps something we will address or what we want to address in 2020.

Very cool. Now I know there's a lot of other sessions in that track. Also customers customers talking about how they are using RAM, so I can encourage everyone to obviously check out the other breakouts in the track as well. Now, you mentioned one thing, I think you called them the properties, right? The session properties. That means for people that are not aware of this, this actually allows you to additionally to capturing the end user in every every single action capture additional business context or maybe data from I don't know I think I

talked with with your colleague or our colleague Klaus earlier right capturing data from other frameworks whether it's from Adobe or Google Analytics I'm sure there's many other frameworks out there so we can enrich our data with this data that gives us more business context and then we can obviously look at the data and data tracing can answer more business relevant questions exactly so what the what it is really about is really on segmentation it's enriching data with really application specific properties this might be a member status but you can also think about using it for performance analysis in order to do A-B testing. So perhaps if you have

your application running on two different technology stacks, what you then can do is add for the new technology stack a specific property and then of course look at each individual user action only for this technology stack and compare it with the old one. Yeah, that's cool. Or, you also said we have a lot of preview programs that we constantly push out there. Preview, I'm sure our customers also use, let's say, feature flagging or, let's say,

canary deployment or an A-B test. So you can really use Dynatrace RAM to segment what type of version is exposed to which user and how does that have an impact, positive or negatively, on the end user so they can learn and therefore optimize the stuff that needs to be optimized. Yeah, you're completely right. And also another big area is of course that in the roadmap session I want to tackle a little bit the mobile monitoring.

So what is coming there? Will Flutter support be coming to Dynadrys, for example? If you are interested in this, please join the session and then you will find it out. Cool. You also mentioned session replay in the very beginning. So for those people that have never seen session replay, obviously check it out. It's an amazing technology, very helpful. Is anything happening on the session replay side that you may not be aware of? Yes, of course.

We have Session Replay now since one year in the product and as always we need to improve. We learn from the feedback from our customers. We also want to make it more data privacy out of the box so we will introduce a new mask all behavior out of the box so that that customers really can be sure that there is no data privacy being recorded by session replay and very important for us as a company of course as a performance company we will add resource delaying, resource capturing to session replay. So that means this is a future project we are currently working on, that in future the replay will show exactly the slow resource loading

what the end user has experienced. Oh, that's awesome. And obviously, because that obviously has a big impact on real user experience. If certain things load slow, that's awesome. Yeah, exactly. Cool. Hey, Alex, thank you so much. I know we have to run to our session because it's a busy event here.

In case people want to find you and have a chat with you at Perform, are you at some of the Innovation Towers? Yes, we have our Digital Experience Management Innovation Towers. I think all of my colleagues, all of my PM teams are there. Reach out to us, ask our questions, ask your questions, take a look at our demos and of course challenge us with new use cases. Cool, awesome. Thank you Alex. Thank you Andreas.

See you later. Bye.