 Hi, this is Yoho Sapil Bharti and we are here at CubeCon in Chicago and we have with us once again Austin Parker, director of Open Source at Honeycomb. Austin, it's great to have you back on the show with a different hat. Yeah, literally changed hats. So it's wonderful to be here. Talk a bit about what attracted you to Honeycomb. That's a great question. You know, it's always really exciting to be at a startup and watch it scale and grow and then get to an exit like we did with LightStep and it's a super valuable experience that's hard to replicate. But there's a difference between being part of a startup and growing it and scaling and then being part of a, you know, much larger company. And I think I had gotten things to where I could get them at service now and it was time to, you know, go back into the startup grind. So I went with, I chose Honeycomb because it was actually really interesting, you know, really interesting story. If you look at the industry right now, Open Telemetry has been this really cool kind of black swan event for of the observability market because it's taken this old model of tightly integrated, you know, tightly coupled vertically integrated tools from visualization, query and storage and instrumentation and throwing it on its side. And now you have instrumentation as this portable base layer that people can build on top of. And so you look around KubeCon today and you see dozens of new entrants that are doing observability tools. And what I saw with all of them is they all looked a lot like Honeycomb. And when you're, when you're looking at they're like, well, imitations and zeroes form a flatter, right? If everyone is copying you, then you must be doing something right. And so rather than, you know, kind of go to something new, you know, very new, it made a lot of sense to go to the people that sort of invented observability and popularized it and have been innovating in this space for years now. The interesting thing with folks like you is that even if you change the company, the technologies, you know, I still talk about the same thing that I was talking about with you for two or three years ago, which is, I keep me, oh, you're here, but we still talk a bit about the kind of evolution that you've seen. Of course, Light Step, you know, there are some of the pioneers, you know, open telemetry, open sensors, a lot of technology came together. But you have seen there, what kind of adoption you're seeing there, and then what kind of role come to that Honeycomb in playing digital? Well, it's been a really exciting year for, you know, open source observability in general, right? One of the things that we've announced as a project is open telemetry this year is that logging, you know, is stable and it's generally available in the specification, which means that open telemetry has achieved the goals it set out for itself five years ago when we started the project. Logs, metrics, and traces are, you know, done is not the right word, but they exist, and the old stuff is deprecated, and here's the new stuff that works. So we've accomplished that original goal, and every year what I have seen, how I've seen that reflect is I come to KubeCon, there's more people talking about open telemetry. I look at big enterprise companies, small companies, mid-market companies, and everyone is really saying one thing, which is we maybe are not sure about what tools we're going to use up at the top of that stack, but we know we're using open telemetry at the bottom. I was talking to a very, very small startup, you know, with some really smart engineers, and they were looking at, you know, what are we going to do for observability, and they actually chose not honeycomb, right? They looked at the options, they said, well, you know, we don't really need this right now, we're very, very small, we're not even public, we want something, you know, we just need these basics, but what they did say is we're going to build on open telemetry, so when the time comes, it's very easy to move into one of those other tools, one of those commercial tools, and that is exciting to me because it means that there are so many people having that same conversation, so many smart engineers saying, hey, if we build an open telemetry, it gives us this optionality, and we are able to, you know, grow into something. After starting out open source, we can move into a proprietary solution like honeycomb, we can make a mix of things based on our needs, and I think that's, you're seeing it at that small level and you're also seeing it at the enterprise where people are going in saying, hey, we're building a strategy for the next 10 years of observability at this bank or at this, you know, big games company or whatever, and they build on open telemetry, so that is the adoption story to me, is it's becoming fundamental both at the big, big fundamental enterprise scale, but also in that small, you know, startup scale. Earlier you were talking about, when we were talking about observability, you were talking about logging, monitoring, those things are there, and if you look at open telemetry, how is the evolution there itself? One of the ways that open telemetry is evolving is as we get more adoption and we have more users, you know, we're having to, we have to listen to them more, right? Like, it's one thing when you're, you know, building something, you don't have a ton of adoption yet, and you just kind of change things willy-nilly, and now what we're seeing is the shift towards like, oh, we have a lot of people using this in production every day, and we have to think really hard when we're about to change something, and that's been a, I think a really, that's been a process for a lot of us as we're learning and growing, and it's also been a process, I think, for the industry, you know, we're at large because you have to change how you think about even little things, like the names of, the names of certain things, because so many people depend on them in so many ways that you can't even see, whereas with the proprietary stuff, it's easier for people to, you know, understand that. So we've had to look at tooling, we've had to look at, you know, how do you kind of solve these modern problems that we have created for ourselves? The other couple things I think are interesting is we're looking, you know, we don't want to just bring in everything, you know, controlling scope is really challenging for any project and open telemetry is no exception, but there's two areas that we're really looking at in terms of growing based on what we're hearing from the community and what we kind of see in the, you know, what people need, and one of them is real user monitoring, you know, and making open telemetry more useful for, you know, web clients, iOS apps, Android apps, things like that, and the second is continuous profiling and being able to get, you know, that really, really, really low level data about what's going on in the code and attaching it to everything else in the open telemetry ecosystem. Are there any specific use cases or industries that Honeycomb is looking at, or it doesn't really matter, whoever is looking at open telemetry, observability, that's very serve. We are an observability company for everyone, right? Honeycomb, we didn't, you know, we popularized it in a lot of ways, we really invented what people think of when they think of observability, I think. For a long time though, people have thought that this is very difficult to use and it's just for experts. So one thing that we've recently announced is Honeycomb for Kubernetes, which is really unique because a lot, if you think about Kubernetes from an observability point of view, Kubernetes emits just a raft of metrics and logs and traces and just so much telemetry data about what's going on because Kubernetes is pretty complicated. And in a lot of organizations you'll see, you know, you'll still see this divide where you have your app devs over here and they are looking at application metrics and application logs and application traces and then you have your infrastructure devs over here, your SREs and they're looking at Kubernetes metrics and Kubernetes logs. But a lot of times, the people in that first group, those app devs do need to understand what's happening Kubernetes in order to understand how things like memory pressure on a node or over under scheduling, over provisioning might be impacting customer experience at the end of the day. So what we have done is we have kind of pulled this curated set of metrics and logs and traces out of your Kubernetes cluster and we present that to you in Honeycomb and we're using open telemetry to power this. So there's a Helm installer that just puts this on there, we have a couple of little things that will get this data and really get rid of this stuff that doesn't matter, right? Because we're really focusing on how do we help people that are developing these applications, building on Kubernetes, building on cloud data, how do we help them understand the relationship between Kubernetes performance and their app so that they can find problems faster, solve incidents quicker, reduce TCO, all this good stuff. You talked about Kubernetes complexity and that is a well-known issue there. There was a few weeks ago a thread on Twitter as well, what's next? Since you have been in its industry ecosystem for a long time, what are you seeing? First of all, the fact is that Kubernetes is now the kind of technology which is like Linux kernel that is done, it's not the A, this is going to replace X, Y, Z. So that is not the question. The question is to make things easier like Linux kernel, it's complicated. Like kernel maintainers, they are not everybody can go and become a kernel maintainer, but there is a lot of ecosystem to make it easier for leverage. So what do you see there? So it's really interesting that you bring up Kubernetes because I was talking to someone this morning and if you think back like six or seven years when Kubernetes was really starting to get a foothold, there was a lot of companies that came up with Kubernetes installers and everyone had their Kubernetes installer, their own opinionated, let's say you do Kubernetes. And now, what was the last time anyone thought about a Kubernetes installer? A lot of the good ideas got put into the project themselves, but the market changed too. People that are running Kubernetes on their own hardware are often using something that gives you a lot more than just Kubernetes, like an open shift or things that pull in a lot of different extra things for value like improved RBAC or various sort of provisioning and security things and mutual TLS or whatever. There's so much that you can put on top of Kubernetes, but fundamentally I can still take that application that I wrote for the Kubernetes APIs and install it on OpenShift or I can install it on Digital Ocean Managed Kates or in Google Kubernetes or in Amazon Kubernetes or Azure Kubernetes. And it's that API that is the standard. So what I think you're going to see is two or three things and it's actually shared between open telemetry and Kubernetes. Because what both of these projects have really done is defined the API that people use to interact with the cloud native world. And open telemetry defines that API and it defines the data model that people use to express telemetry about their services. And in both cases, what I think you're going to see are one, these APIs becoming more and more standardized and built into more and more things. Even things that don't necessarily look like Kubernetes, maybe even the things that don't look like open telemetry, but those APIs will be the way that we interact with, you know, as developers that we interact with those telemetry subsystems or those container orchestration subsystems, right? And the data model will be sort of the standard default that people use to express these telemetry points. And that's going to require changes, I think, in a lot of the market, like a lot of the ways that we think about business models, especially around monitoring. It's like, well, you have this many hosts. Well, what's a host in Kubernetes, right? Like, what do I actually get by monitoring every node? And my nodes can change so much. It's going to be a really interesting few years. Austin, once again, thank you so much for, you know, getting time out and talk, you know. And I would love to chat with you, but not only so much is happening in the ecosystem, but you have so much insight. So I look forward to our next discussion already. Thank you. Absolutely. It's been great talking to you. And I said, I can't wait to talk again in a few months about everything. It's going to be surely all different then. Thank you. Cheers.