 Welcome to Amsterdam and KubeCon, CloudNativeCon 2023. Join John Furrier, Savannah Peterson, Rob Streche, and UPSCOT as the Kube covers the largest conference on Kubernetes, CloudNative, and open source technologies together with developers, engineers, and IT leaders from around the globe. Live coverage of KubeCon, CloudNativeCon 2023 is made possible by the support of Red Hat, the CNCF, and its ecosystem partners. Hello, we're welcome back to the Kube's live coverage here at KubeCon EU, also CloudNativeCon Europe. I'm John Furrier, host of Kube with you, Pizari's with me. Savannah Peterson, Rob Streche, all here, breaking down all the action, three days of wall-to-wall coverage, got great leadership interviews with company CEOs, topics, we're expanding on new format with panels, and just unpacking key areas. We had AI earlier, we just had the future, younger generation, and diversity. This section's about observability, the hottest sector I've been watching with the team for many, many years, hype cycle now is relevant, got a great panel here, Mike Kelly, the CEO, ObserveIQ is here, Mike, thanks for joining us, appreciate it. Oh, thanks for having me. And Eduardo Silva, CEO of Boundary Clip, yeah. Guys, you're leaders in observability, thanks for coming on. That's great to be here. Thank you. So, observability's gone through, evolution kind of mentioned it at the top, you've been talking about it, like it was really hyped up, a lot of funding, a lot of acquisitions. Some companies dropped away. What is observability? That is the biggest question. Yeah, yeah, there are many answers to that question. But there's a technical answer, and I think the technical version is something like, it's the ability to know the state of a system, right? That means it's observable. The pragmatic, it all lets you jump into it. Yeah, pragmatic technology will be like, you need to know what, how your applications are behaving, and actually one of the observability exists because you want to analyze your data. It doesn't matter what is the form, but you want to do analysis. And to get there, you know, we get observability, before that we got different names, now we have the concept of telemetric pipelines, and so on. It's very technical. We always call it plumbings, all kinds of definitions, softwares involved, data. We're going to get into the AI discussion, kind of where observability's going to go. But first, before we get started, give a quick minute about your company, what you guys do. We'll start with you. What's your company do? What's the purpose? What's the North Star? Yeah, so, yeah, the company is ObserveIQ. We're focused on open source observability solutions, and specifically really on the telemetry layer. So how do you gather, transmit, transport, and filter out data? So observability pipelines are really what we're doing. We do a lot of work in open telemetry and have released the product by and plain OP. There's an open source version of that to allow you to manage agents at scale. And open telemetry has actually been doing really well on the project, quick up straight there. Quick update on the project? Yeah, it's been, so, big announcement just this week was the elastic schema is getting merged in with open telemetry. I think that's a great step in the right direction. Anything that can kind of combine these standards makes it easier for everyone. And we've been seeing that across the board. I'd say also, there's been a shift over the last year that we've really noticed with open telemetry where this vendor supported, which is great, but now we're seeing all the customers are really embracing it. And that's been a huge benefit. We'll come back to that. Eduardo, what do you guys do? Talk about your company. We are Caliptia. You might know us before we were maintainers and creators of Fluent D and Fluent Bet. So we started this journey two years ago, but in the observability space today, we started 10 years ago building all these documentation pieces. And now from the company perspective, we're bringing the same solution that we built for the community, but for the enterprise, right? Right now, there's a huge problem where people is generating more data into their system, but value doesn't correlate. Actually goes down. If every year you get 20 to 30% more data, you don't get the same value. And actually we found that in observability space, everybody was focusing on the storage of the data or where they're running analysis, but nobody focused in the previous part. We called it the first mile. So Caliptia, we focused on the first mile telemetry pipelines or call it observability pipelines. So what I'm interested in hearing is, I mean, you raised an interesting point, right? You said that last two years feel like 10 years. And that means the space is evolving very quickly. What I wonder is, from a customer perspective, how has that changed? Because I know observability from two years ago, I know it now, but how has it changed from a customer perspective? Because I think the use cases have evolved rapidly. Yeah, actually we can give you a quick overview. The cases from 10 years ago was different sources, different formats of data, but they still need to do analysis. So the problem remains. Now, if you fast forward, I have more sources, more formats, but at 10 or 20 X scale. And that keeps growing. And now we are in a stage where people want to standardize on type of telemetry. I would call it like, now we have diversity in observability because the diversity now is logs, metrics and traces. Everybody just thought about logs before, right? We have monitoring system for metrics, but now it's different. So the way that, if I go, for example, I refer to my customers right now, one of the problems that they have is performance, right? They cannot move the data fast enough, or most of them is like, oh, implement your solution and now we are getting more data. Yeah, because you are not collecting that data. So the scale is quite high. The other problem that exists right now is like with this telemetry, you need to integrate and connect different systems, right? So most of companies are in a journey on this to have a vendor neutral approach to observability, to not get into a vendor locking by vendors. An open source aspect of this is becoming a big part of it. We're seeing more focus around developer productivity. What's changed now? Yeah, that's been interesting. There's been a number of, I think, critical projects. Fluent D, Fluent Bed are certainly massive and open telemetry has been a big one. Prometheus and others in the space. And they've all been really focused on standardizing that telemetry layer. And so when we're thinking about that, we've been talking about telemetry, and I completely agree. One of the biggest challenges people don't recognize and it's more and more of a challenge over the last few years is the quantity of data is just overwhelmed. You can't, customers can't ingest and manage that data even if they have the analytics support and the platforms in place to do it. And so finding a way to, number one, standardize that layer and then reduce data as you're collecting and shipping it is critical. And that's where I think things have changed. And the complexity is getting, like you mentioned, the complexity is higher. All right, so now let's zoom out. AI is upon us. Everyone in their mother's used chat GPT. He goes, oh my God, it's magic. It's like in a Harry Potter movie, you know, like, you know, Stoley Armist. You know, things just magically happen. One of them is auto-generating code. So we are like seeing, if that's now a new factor, that's data. So you got a ton of data. Now this community is, I wouldn't say they're very data-centric in the sense of they don't talk about data all the time. Data is super important in observability, but like now AI seems to be coming fast here. Where do you guys see the impact of AI? I mean, I think it's, no one knows what the full impact's going to be, but it's already clear that it's going to impact every industry and every company in one way or another and you're either on board and using it or you're going to be behind. I think when you look at observability, it's really interesting because AI has always been maybe a buzzword that was put in with a lot of companies. It didn't necessarily, you know, I think a lot of customers were feeling like they were let down by that in the past. But I think now we're at a point where it means something and you really can't accomplish a lot with it. So just to give a simple example, right? In the past, if you think about integrations, if you wanted to get data from a specific application, you would need to write something that would parse that and tell you, how do I interpret that? You can trivially put that in chat, GPT, and it'll tell you how to parse it now, you know? These are things that we can do very easily with AI today and that they're likely to change the industry and certainly make tools more productive, people more productive. Because one of the issues with observability is being able to interpret the data. So I think there's a lot of manual work that still goes on in that space. So now we do have, you know, we have a simple platform. We have observability pipelines. A lot of the boilerplate has been taken care of, but we still have to interpret whatever's happening to take meaningful actions. Because in the end, this is all about performance of the application or stability or cost optimization or the list goes on. But it still requires us to optimize and interpret that piece. So what I wonder is, what does AI look, you know, what will AI solve for in that sense specifically? Yeah, yeah. Actually, we have some experience because we were, the company started experimenting with this. If you look at the problem that's limited pipe and solves, it's about to give the user control of the data. When you give control to the user, the user had to choose path from source to destination. And there's one thing that adds a lot of value, which is allow the user, customer, to bring the business logic into the pipeline. Business logic, so we're chiefly left, right? So when the user is going to start taking control, I want my business logic, they always come out with this question of, okay, I have a custom business logic, I'm flowing, for example, credit card transaction number, transactions in general. But I need a processing rule before to send my data that does XYZ busy. Okay, so from a product perspective, we allow the customer to do some scripting, some stuff, but we found, hey, the problem is that the user needs to get a way to simplify this process. And what we did was an AI implementation where you can write in human text, hey, text my data that looks like this and do XYZ. Click, and the AI will generate the processing rule for you. I think that we're in the face of simplifying the control of the data, but it will take maybe a few years to get to the AI discovered. Yeah, I mean, you worry about the hallucination aspect of a chat GPT, now the one of the worst, not worst, but like one of the most over the top things I've heard this week was other entrepreneurs, their investors say to them, now this is not the entrepreneur, this is the investor. What's your chat GPT strategy? They're in the observability space. Like, that is the dumbest question. Now, what they really mean is, what's your AI strategy? And are you going to be on the right side of history on this thing? So I don't think there's a definitive answer. It's more of, are you going to be on the right side of the AI benefit, which is augmentating the humans, having some operational benefits, not getting lazy and leaning on code. So I mean, it's all kinds of, it sounds a lot, but it's not yet clear. Yeah. And I think of it as a few different components. And if you broaden it out to what a company needs to do to adopt it, there's a component that is what are the tools that you're using. And right now, Copilot and other tools for just development have really increased productivity for developers. But then there's also, how do you develop processes with AI? And those have even more significant, I think, longer term impact. And then how does that integrate into your products that you're developing? Because I think it's going to impact all those three pieces. And for any company, it's kind of like making sure that you're doing all those or at least taking advantage of as much as possible. And what's your thoughts on this piece? Yeah, I think that as, well, in the AI space, everybody needs, is very concerned about security or all aspects, auto-generated code, right? But I think that we have to go beyond that and always think on what is the user experience and how we can use UI to improve that experience. And as I said at the beginning, how we can extract more value from it. Yeah, yeah. So I think, yeah, we are in a very early stage of this and yeah, we're looking for what is not being the wrong or the right side, but I think use the right tools. No, but there's also the aspect of security, right? And you mentioned this. There's the aspect of the data has to go somewhere, your data may leak, like AI poses a risk, right? On the other hand, having intelligence that helps you prevent leakage, helps you prevent having all of your data sent to another AI. Like there's both a risk and an opportunity. And what I wonder is what's that development going to look like? What's that feature going to be? Yeah, yeah, and that is one of the things, a good point that you bring up which is the security concern. And a lot of times when you're dealing with this type of data, it's not something that can be passed along to a tool like this, right? Because that could be considered a leak for sure. And so that restricts what we can do. And it does mean that there are new models that need to be created to even take advantage of, you know, the new large language models that are out there. What are the big customer, developer, consumption side of it? Developers are actually using open source, obviously a big part of it. As they commercialize with telemetry and observability, where are the customers right now on this? Because there's a lot of architectural stuff going on. People selling repatriation, we're going to go on premise. That to me, I think it's just more about cloud operations. All right, okay, I get that. You got edge coming around the corner, AI is now booming, people see the role of data. What are customers doing? Well, you guys have customers dealing, what are they thinking about? Where are they in this picture? Are they kind of in early stages of school learning the alphabet? Or are they further along? How would you describe them? With AI in general. Just observability and getting the right telemetry and data monitoring down. I think a lot of, this is where we've seen a lot of teams that feel like they've got it figured out. They did, they said, all right, we have our security stance, we have our observability application monitoring is in place and we're doing it well. And then it got out of control. But that was where it was, there was too much and now everything was kind of crumbling. And so they're trying to figure out, how do I keep what I had a few years ago with 10 times the amount of data that's coming in? That's, I think there's still a lot more and where you see innovation happening right now tends to be on, how do we take that, not eliminate the visibility, but somehow either compress or find new ways of evaluating the data. Edward, what's your take? You talked about value earlier, customer's money and the value side of this. I think that the problem that we have 10 years ago, we still have it. Yeah, and most of the customers that, the problems that I see, it's about scalability, right? That is what is one of them. The other is unifying this old type of telemetry or observability data with new platforms. For example, we got the, so from the open source ecosystem with flow and bid, everybody's deploying or we got banks that have deployed 100,000 servers with the agent that didn't have the agent. Now they want to have a more vendor neutral approach to observability. If you ask why they need a vendor neutral approach is because now they want to be diverse also in the platforms that are used for observability. If you look at five or 10 years ago, they didn't have that option. They have one vendor and that's it. Now the user want to have control, the user want to have a way and a vendor neutral strategy to move forward. Yeah, I mean, we heard this with WebAssembly yesterday, we had a big session on unpacking that and we were discussing like the data mesh area, I don't know if you guys follow the data mesh, data bricks, no flake of the world, they have open source vibe, but their tool chains are proprietary. So now we're getting into this, okay, to your point about if this continues to grow, where's the clients on the lock-in side, where, it's a tough, because you want to have the best tool but the platform engineering is rising up to be that new layer of kind of stability, what's your take on that? Are my missing the boat here? You always see that balance, right? There's going to be a push towards open source and open standards. I think the key here is in the open standards, right? So are we sending data in an open way using open formats? Because that's what prevents you from getting locked out and one of the challenges now is it's very difficult in observability to shift, to try new things because you put in, there's a high cost instrumentation, there's a high cost end vendor lock-in and that sets people back. I think that people are recognizing we want to be able to switch, to be able to try different tools and then iterate quickly. And if you can do that, I think it proves out to be really effective and you're going to be much more productive as a team. Eduardo, what's your take on that? I mean, these guys are in the middle of it. Balance, open, you got to make money, but where's the differentiation come in? Scale obviously is one competitive thing. I think that without open standards, there's no way we can scale in the future. That's the fair thing, but also we have to acknowledge that changing patterns and architectures in companies, it doesn't happen in one day. And standards takes time to be real. Standards in production. It's for example, yesterday we were in a session in Flonbete, people was asking me, hey, open telemetry versus Prometheus. I was talking about locks. And I told them, you know, yeah, open telemetry is a standard. And open tele, pardon, Prometheus is what the industry is running for metrics. So, but it's up to you. At some point they will be achieved on these alignments, but what we need now is some vendor neutral strategy to unify all this movement. You cannot switch from one to the other, right? And actually I was telling, yeah, who has a company? Yeah, are you running Oracle? Yes, are you running POSRE? Yeah, and MySQL? Yeah, at the same time, the same thing happens in observability. Yeah, and I think that's the key. Let's get in the market. It'll work itself downstream. All right, before we get to the final question, I want to ask you guys as CEOs, it's a great market. I mean, there's a lot of dynamic shifts in the technology. You see a lot of growth ahead, tailwind in this market. This is 10,000. As you guys look at the landscape, what are you guys thinking about right now in terms of where the observability market's going to go? What's your bet? What's the 20-mile stair out there? What are you prepared for? What do you think's going to happen? We'll start with you. Yeah, I would say that the first thing is everything is shifting to the left, the problem. So I would say that the focus of observability will mostly in the first mile, by controlling the data where data is being generated, that is the trend that we have seen in the open source and now with Calypcia in the company. And well, as a company, we're focusing our solution on that place and be a best friend with every single vendor. Yeah, so I think that there's an emerging need for a different layer, and that is the ingestion and management of the data, the telemetry platform. And that feeds all of the analytics, your security and your observability solutions. I think that that's going to start to become, start to gain momentum. We're going to see that really as a need for most companies. And that's a big shift for folks. I also see that there's going to be a combination of business analytics and what we think of as IT, observability and security. Because it really, it becomes a lot of that same data. And right now there tends to be a business wall between it, not a technology wall between it. Yeah, that's a good point. I mean, we were talking about the developer. Yesterday we were riffing like, who makes the decision to store the data? Not developers. Like, what if that flipped around? What if developers can control how the data's managed so that they could program it? So, you know, this is where if you get access to better telemetry, better data, I just think this, it just feels like the world is about to just spin in another direction. Well, and I think we're kind of flipping it on its head because right now it still feels like we're solving an organizational solution with an organizational problem with a technical solution. And it feels like we're on the cusp of switching that over, that we're actually going to solve that organizational problem, not necessarily just with technology, but moving it over into a broader sense where we solve the organizational issue, but not with technology, but we solve that part and then we can enable the technology, we can fully leverage it. And like you said, we can then remove that wall in between it. I think that's a fair point. That is going to enable a lot for us. Yeah, I would say that this is a really interesting journey and you know, in observability and anything that he can agree with that, I think that our proposal, even as a company is from what I know from his company, it's like we don't try to be a dropping replacement for anything, actually most of some companies say don't use Splunk, don't use XYZ because they're really expensive. And I think that the right approach is from a observability perspective, hey, let's optimize so you can get most value from your backend and your technical decision. I think you guys are right on. I think that also the observation is there are scenarios, things pop a certain way that could fall different paths. One path is data-driven. That's AI, fertile ground for machine learning and programmability. And so that's a big thing that we're watching. Security, I think, is easy. It's going to be bad and good. I think it's going to be interesting use cases coming out of the gate that's going to be malicious and offensive defense, so all good stuff. Well, guys, thanks for coming on. I really appreciate the expert advice and commentary. But before we go, give a plug for the company. We'll start with you guys. What are you guys up to? You hiring? Yeah. How many customers, funding? How many people you got? What are you doing? Definitely hiring. So looking, if you're in the observability space, love to hear from you. And yeah, as a company, we released Bind Plane OP recently. It's a platform for managing your telemetry. It manages open telemetry agents and that helps you to reduce and filter out the data you don't need and send it to the right locations. Awesome. At Calitia, we are a 30 people's company. Calitia has been around for two years and today we're launching Calitia Core, which is primary product version 2.0 that comes with aggregation, processing rules to bring you to business intelligence into the pipeline plus fleet management because we believe that also, all these telemetry pipelines needs to be distributed, not just centralized. All good on the funding. You guys need funding? Venture Capitals are watching. For now, you're good. Okay. Yeah. Good. Keep that runway going. Yeah. Well, luckily you're on the good side of the market here. It's been a really kind of a downside on some SaaS, some other companies might not make it to this market. So congratulations. Gentlemen, thank you for coming on. I appreciate it. Thank you so much. Okay, for Yuppa and I for a year. Thanks for watching. We'll be back with more live coverage day two. KubeCon EU, we'll be right back.