Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Building Grafana Labs and the Future of Observability with Anthony Woods
9 Plays5 minutes ago
In this episode of the Kubernetes Bytes podcast, Bhavin talks to Anthony Woods about all things Grafana Labs, Observability, Telemetry, and how AI impacts both of these ecosystems. The discussion starts off by talking about the early days of Grafana Labs, what is Adaptive Telemetry, and how AI plays a role both in building Observability capabilities in applications, and how it helps perform root cause analysis. Listen to learn more!
Check out our website at https://kubernetesbytes.com/
Show Notes:
- GrafanaCON 2026: https://youtube.com/playlist?list=PLDGkOdUX1UjoSfz1IRj5c0xetw8tl8iin&si=JpT85m4t4bP8ZXgX
- Grafana Labs Blog: https://grafana.com/blog/
- https://www.linkedin.com/in/anthonywoods1/
Recommended
Transcript
Introduction to Kubernetes Bytes
00:00:03
Speaker
You are listening to Kubernetes Bytes, a podcast bringing you the latest from the world of cloud native data management. My name is Ryan Wallner and I'm joined by Babin Shah coming to you from Boston, Massachusetts.
00:00:14
Speaker
We'll be sharing our thoughts on recent cloud native news and talking to industry experts about their experiences and challenges managing the wealth of data in today's cloud native ecosystem.
00:00:30
Speaker
Good morning, good afternoon, and good evening, wherever you are. And we are coming to you from Boston, Massachusetts.
Episode Date Announcement
00:00:35
Speaker
ah Today is May 29th, 2026. Hope everyone is doing well and staying safe.
00:00:42
Speaker
I guess listeners in the US, hope everybody had a great long weekend for Memorial Day. ah The weather in Boston was not great. um I think it it rained over the weekend, but I was able to fly out on flying day to a Friday to warm and sunny Arizona, which I was glad for. The weather was warm, maybe almost too warm. I was not ready for 90 degree weather, but it it was good to be out there. um But I'm glad to be back and excited to bring you guys another episode.
Interview with Anthony Woods Preview
00:01:10
Speaker
umt For today, we have a great episode lined up, an interview with one of the Grafana Labs co-founder, Anthony Woods. And we'll ask him everything about how they started Grafana Labs, how AI is impacting the observability and telemetry ecosystems, and what tips he has for everyone or anyone who's looking to start their own open source
Anthony Woods' Background
00:01:30
Speaker
project.
00:01:30
Speaker
So with that, let's get Anthony on the show. Hey, Anthony, welcome to the Kubernetes Bytes podcast. It's great to have you on the show. Can you take a minute and introduce yourself um to people that don't know you and then talk about what you do?
00:01:44
Speaker
Sure thing. thanks for Thanks for having me here today. and So my name's Anthony. I'm one of the co-founders from Grafana Labs. um So I'm a you know technology enthusiast. I always have been. um Interesting, I live in Perth, Western Australia, beautiful city, just ah a long way from everything. But that's okay in a post-geographic world. It's part of why Grafana Labs as a company is a fully you know remote company, the three founders. We couldn't have been further apart if we had tried when we started the
Journey to Grafana Labs
00:02:14
Speaker
company. And so we've always been you know been focused on being remote and and building out the company that way. But as I said, I'm a technology enthusiast, um you know started my career as a network and systems you know admin and then you know worked a lot across the technology stack around storage, infrastructure, networks, you know you name it. And then moved into building more and more kind of glue. So I started off with some hacky little scripts. And over time, that grew and grew and grew until one day I found I was just writing software. um and And that was a lot of fun, right? Being able to kind of work in the infrastructure space and and build great tools.
00:02:46
Speaker
And yeah i was I was lucky to ah to you know meet Raj, our CEO, um when I was living in Singapore. he He had a company he'd started in his dorm room. And um yeah it was a great opportunity to work with him, get to meet him as a great person. But then also, yeah i got to work in enterprise work at Visa. And I was there when I discovered how horrible enterprise software was and the fact that I could replace it, right? So,
00:03:08
Speaker
um That was a great inspiration and and why I wanted to go and do a startup. So, you know, that led me back to Raj and, you know, we wanted
Formation of Grafana Labs
00:03:16
Speaker
to work together. And then ah we met up with Torquil around the Kifana project. So it was a great... um You know just kind of things just kind of aligned. Right. And we kind of just found each other um and and we got got together working on building, um you know, what is now Front Labs. Right. Starting with the open source project. And yeah we had a vision of build great open source technology and we'll work out how to make money later. um And that that was the original business plan. But, ah you know, things have obviously developed since then.
00:03:41
Speaker
No, that that's awesome. Like, it's interesting. So were you and Raj in the same location? Like both of you were in Singapore or that was distributed as well? That was still, even though, so yeah he'd started ah his previous company, a company called Vox, out of his you know college dorm room in New York and um you know grew the company in New York, but wanted to expand into Singapore to have better time zone coverage. And so I was the first employee in Singapore kind of you know built out the team there. And you know obviously he'd pop in very frequently, but it was still, even then it was a a remote company um yeah know with half in, in US and half in in Singapore.
00:04:13
Speaker
Gotcha. So like the interesting comment that you made, right? Like working at Visa, giving you the idea that, yeah, I can do better. Like we can build better solutions. Go ahead.
00:04:25
Speaker
i was going to say, yeah, was actually really interesting because you know I'd, know, I've been working, um you know, for a lot of different companies, you know, throughout my career and and a lot of it was around me and, you know, working with teams and us building our own, uh, technology. You know, that was actually a great thing, right. In hindsight, right. The people that I was able to work with in my early career were yeah really talented, you know, highly skilled people. And I had great opportunities to just,
00:04:44
Speaker
develop my skills, right? And learn new things. Um, but you know, you, you always think the grass is greener on the other side. so I was like, Oh, maybe it'd be interesting to go and work in enterprise. Um, you know, not have to build my own tools, be able to use these, you know, enterprise software that they talk about. Um, and yeah, just discovered how horrible it was. uh, and I was very fortunate, my manager, you know, at visa, you know, gave me the the freedom to replace it. I remember just being fed up. I think it I was only there for like a month or two when I was just like, this software is terrible. And I spent a weekend and built a proof of concept to replace one of them and showed it to my manager. And he was like, yeah, you can just spend your time doing that from now on. I was like, okay, that sounds fun. Yeah.
00:05:18
Speaker
That's awesome. Like this was, this was before wipe coding and people building apps quickly in a weekend. So that's impressive. Anthony. Yeah. We can do like cobbling together existing open source technologies and and then, you know, building on top of it. So.
00:05:32
Speaker
Yeah. So what was grafana the first thing? And was it open source from the beginning? Like you just created a good repo and got started with it?
Origins of the Grafana Project
00:05:40
Speaker
Yeah. So it actually, um you know, Grafana actually predates, I guess the, the company, right. So Torkel, um, you know, uh, you know, the, our other co-founder is based in Stockholm. He created the Grafana project, right. Um, uh, and like all the projects, you know all the products we have at Grafana Labs, it's it was really just around like scratching his own itch, right. Where, yeah He was working as a consultant at eBay Sweden. They had a lot of data. They were collecting kind of telemetry out of their applications, but nobody was looking at it or using it. um And his hypothesis was that it was simply because the tools to build dashboards were just terrible. and He thought if he could build a better dashboarding tool, um people would use the data. um Turns out he was right, um right? Once he built Grafana. it had immediate um adoption um you know and it kind of spread like wildfire and and part of that is you know teams would get it they'd they'd hear about it but they'd you know they'd build a dashboard for their team they put it on a tv in their office and then other teams would walk past and they'd see that dashboard and go that's amazing i want that too and so then just have this like viral spread of you know grafana through organizations Um, and it was super popular. Um, and then that was in the very beginning. So he launched that project, you know, January 1st in 2014. that's when he kind of shared that with the world. Um, and then Raj and I met up with him around October, 2014. So, And Toggle had already, you know, he quit his job. He'd seen the, you know, the growth of Grafana and he's like, yeah, okay, this is opportunity. So he was working on it full time. He wanted to build a business about yeah around it, but didn't really know what what he needed to do to make that happen. And Raj and I, we'd we'd gone out and we'd say, hey, we're going to do a startup together. We want to do something that was SaaS based, that was, in the kind of monitoring observability, uh, kind of space. And we were like, we saw Grafana and we're like, Hey, this is great technology. We could use this to, you know, jumpstart, you know, the product ideas, right. We're not going to have to go and build our own know dashboarding tool, you know, but but we can leverage this existing open source technology. But the more we played with the Grafana, the more we got excited about it. And so, um so we just reached out to Talkall and said, Hey, you know, we love what you're doing. We would like to pay you, as a consultant to come to New York for a week to work with us, to help, you know, better understand, um, you know, the Grafana, um, uh, you code base and technology and how we can integrate with it. But we didn't do any work when he yeah showed up. We just kind of got along, um, really well. You know, we had ah ah a, great shared vision for what we could do with Grafana and a great set of complimentary skills. Um, and so, you know, during you know our week of, of, uh, getting to know each other in conversations, it just made sense that, you know, um, you know
00:08:01
Speaker
we could just you know make the the company about Grafana. So we invited Torqu to join us as a co-founder and we said, hey, we'll just you know we'll pivot and and the company will be about Grafana and that's what we'll go and focus on. And so that's what we did. um And so we focused on just building great open source technology.
00:08:15
Speaker
um And we said, you know our plan was, you know we'll we'll build great open source technology, we'll build a great ecosystem. And then, you know, we'll work out what to do, you know, as we go. Right. And that's you know what we did as a company, right? Like we we got out there into the ecosystem. We heard from what people were doing, what challenges they were facing. And that's what led us to you know building, you know ah you know, what we have today, which is a more kind of bespoke, complete ah observability solution.
00:08:39
Speaker
And how are you connecting with these users? Right. Like is, was it just through GitHub issues? I don't know. twenty fifteen Was there a discord that you guys were running? No, it was definitely mostly just GitHub and then a lot of conferences. So we'd go to a lot of open source conferences, um, you know, whether it was like monitorama was a big one we'd go to in Portland. Um, But there was a lot of you know open source conferences. I mean, it was predates you know KubeCon and those kind of things, but um yeah there were there were plenty of ah of conferences where we could go out and meet um you know meet with the the users right in the community that we're using and loving Grafana.
00:09:11
Speaker
yeahp That's awesome. So did did you guys ever have to um figure out how you can help improve the signal to noise ratio? like Again, you can throw a lot of data, every every component of your application can... can have telemetric data that it sends up.
00:09:26
Speaker
What's actually useful? Is that on the user or you guys also had to make some design decisions and choices? Yeah, we've definitely had to evolve that over time. I mean, certainly when we got started, you know, it was the measure all the things kind of mantra that was going around. and And, you know, obviously that's come back to bite us all with loud there's just too much data. One, it's expensive, but also just when things go wrong, you just don't even know where to start looking. um And so, you know, we've gone on that journey with our customers, you know, and our community, our users, you know, over time. So first it was just, there was, there was too much data um for existing, um you know, databases to deal with, right? So, you know, Graphite was the first one that we started with. And then, you know, we'd have users would come to say, our Graphite clusters just can't scale, right? Like they just keep falling over. um and so you know we set out on a path to build better databases right so we built a more scalable kind of database that were designed for the volumes of data that we're now seeing right and a lot of that is just from you like the rise of you cloud native the rise of microservices right just the yeah the amount of data that people were collecting just kept growing and growing um and so we set out to build better databases um we started with the metrics then we built a logging one then we built one for tracing we've got one for continuous profiling etc um And then over time, you know, you know we were really happy with that. We thought, hey, we've built great databases. um You know, that solves everybody's problem. And, you know, we're very focused around the open source community and that's what they wanted. They wanted a bunch of tools that they could go and use to solve problems. But then as we grew as a company, we started to, you know cross that chasm and move into more and kind of mainstream users. And they they didn't want tools, right? They just wanted solutions, right? So we we then went down the path of like, okay, well, we'll go and build, you know, more and yeah know opinionated, you know, observability solutions that solve a problem, you know based on our experience, you know, based on what we're hearing from the community, let's just bake in what's the the best way to go and do this and and build products around that. And that's what our Grafana Cloud
Challenges in Observability and AI Impact
00:11:16
Speaker
Platform is today. It's really focused around just solving observability problems without you having to make choices about how should I go and about and do this? um And now we've come, you know, that evolution, right, where the volumes of data just keep growing and growing and growing.
00:11:29
Speaker
And so it's both hard to find, you know, the signal in the noise, but also it's very expensive. And so, you know, we're just finding that the cost of observability just kept growing, but the value had plateaued, right? And so we set out on a path of how do we realign the the cost and the value, right? and Just because, you know, certainly as ah as a you know a vendor, right, like you don't want to be, you know, selling solutions to our customers that don't deliver value, right? Like they can't justify that spend. And so we don't have to do it, yeah. Exactly, yeah. They become unhappy um and they start looking for for alternatives. And so, you know, we have a ah part of our platform we call our adaptive telemetry, right, which is really just around, you know we analyze the data, right, so we look at what what we're ingesting and then we look at what you're using, right, what dashboards are wearing, what what alerts do you have configured, you know, what ad hoc queries are your users running, and then be able to analyze that and build, you know, an understanding of what data is actually useful to you and what data is not useful, um and then be able to
00:12:29
Speaker
either discard it or aggregate it away. So you still have some information there, but just be able ah to kind of reduce the the total volumes of data. And so that's been really effective and and it's really helped our customers, you have that control around, you know, that lever to to dial and manage the cost and value out that's right for that that business. And it's actually really, it was great when we explained this to our board and we said, hey, we've got this great feature. It's going to help our customers reduce their bill by 50%. And they were like, this is insane. You can't do that.
00:12:55
Speaker
and And we're like, no, it's a great idea. be fine. be fine. and And, you know, the reality is that, you know, there were certainly some customers who have just, you know, been able to reduce their bill. You know, they're under pricing pressure, right? Like, and that's great. We've been able to help them. We've been able to retain them as customers, right? Like that's really important to us. um But for a lot of customers, was just, you know they have a budget for observability and, you Now suddenly we've reduced their costs, so now they can go and do more, right? Like maybe they haven't started doing distributed tracing, so now they had the opportunity. they're like, well, I've got the budget for it. I can now start using other parts of your platform um you know to to get more value for their engineering teams. And so that's been really a really you know positive thing for us to be able to... any
00:13:36
Speaker
You guys had like Javon's paradox before all this things that yeah, if if the models get cheaper, more people will use them. You guys proved it even before that. That's awesome.
00:13:47
Speaker
Yeah. Yeah. and so that's worked out really well. And again, it just comes down to you that focus for some users. It's just about cost, but for most values, it's really just about bringing value up. Right. Yeah. No. And having done this almost for a decade, right. At this point, what, like, can you share like some of the mistakes that that you saw teams and users make like initially and now right given that observability as a technology has crossed that chasm what are some of the recent mistakes or or errors that you see maybe developers designing telemetry incorrectly like what are some of those things that you you would recommend people like stay away from
00:14:24
Speaker
Yeah. I mean, it's, it's an interesting problem. i mean, definitely observability. It's not a destination, right? It is a journey, right? Like there's, so there's always, um, you it's a path that you're on. you're always going to be making refinements and making, uh, improvements. And I think, and as long as you're, you know, on that path and you're looking for continually getting better, um, then, then that's the right thing to do, right? Like you don't need to you know aim for perfection, um, because it's unattainable, right? You just need to look at where you are today. and and how can we get better over time and and improve things?
00:14:53
Speaker
um I think one of the things that's become, you know, really beneficial, I think, within the ecosystem is is open telemetry, right? So having, yeah you know, open standards are great for everybody, right? It's great for, you future-proofing your technology, right? Like the last thing you want to have to do is every time you want to change your observability vendor is have to go and re-instrument all of your applications, um And so being able to you know leverage open telemetry to you know focus on how do we collect the data?
00:15:19
Speaker
yeah It's got very great consistency, um you know which is great for the users because you know every different team to you know might be using different software languages or they you know might do things slightly differently. But you know open telemetry is a standard and and having great and kind of semantic you conventions around how things should be named and consistency means that you know when teams new members come into your team, they understand what the data actually means, right? It looks familiar to them. And then obviously, you know you get that consistency and you get that um you know ah global knowledge coming into you know the best practices of how to go and what are the important signals that you want to go and collect. And so that's been really great is, you know for a lot of organizations, you know, that we go and talk to, often our conversations are really around understanding what is their observability strategy, right? Because the last thing that we want to do is come in and just sell you a tool, um but not align with, you know, some kind of business outcome. um And so we really focus on like, what is that strategy? You know, what is your vision for observability? And for a lot of those, it starts with, yes, we want to get more consistent, right? so they're going standardize on, on Open telemetry, right? Just having that um that open standard is a great way of future-proofing. And then from there is how do we you know, how do we, where do we go and put that data, right? How do we get the best um understanding out of that data what platforms are going to work well for us? And that's where, you know, we typically come in and help customers.
00:16:36
Speaker
Gotcha. And like when you're asking these questions, is the answer like, hey, yeah we have it figured out, like this is our strategy and that this is the piece that we need help with? or is it like, yeah, we don't know. It's kind of shadow IT, right? Like every team does it does it in its own way. They use a different tool. Like do you see a sprawl of different tools as well?
00:16:56
Speaker
100%, yeah. um So we do our um ah observability survey every year and consistently we see organizations around the world. They have many, many observability tools. And some of that is just because you know they've let teams you know have autonomy and go and pick what they want. So different teams have picked different tools. Often it's because the companies have grown through acquisitions and they've brought in you know, other companies. And so there's just like this, you know, range of different tools. There's always lots of reasons for it, but that is always consistent. It's always a a big part of the the strategy, right? that That customers have when we talk to them is around tool consolidation, right? They're like, we have too many tools. We want to go and consolidate and get rid of them. And, you know, certainly, you know, Grafana, you know, from our, you know
00:17:38
Speaker
just how we've always evolved the company, right? Really focused on integration and interoperability. That was the, you the power that Grafana had from day one was being able to visualize and understand your data no matter where it lived, right? You didn't have to move your data into one system. And that's still true today. And that's big value that we have where, know,
00:17:53
Speaker
people have all of these tools. And so the first step is like, well, how do I get more value out of the investments I've already made? And that typically is like, well, bringing Grafana is that, you single pane view or that first pane view of, of being able to have a a good understanding across the entire, um, you know footprint of your environment, um, and stitch all the data together. And then from there, then they might start replacing, you know, some of their, existing tools with, you know, with, with solutions that we can provide, right. To help reduce costs or to give them better visibility. Um, But often it's the the starting with ah how do we how do we get you know tool consolidation? How do we reduce the number of tools? And and the motivation is just to be able to have a good understanding of what is in your environment. you know I was talking to some customers this week, um and that's ah a problem they're facing today where they've got um you know a bunch of different teams that are using different tools. And so whenever they have ah a major incident, Um, you they don't focus around their meantime to resolution or meantime to detection.
Tool Consolidation Issues in Organizations
00:18:45
Speaker
It's all about meantime to innocence, right? It's like, it's like everyone's just trying to say, it's not my problem. Right. Um, the reason they have, yeah, they have to do that. They have to go and talk to every individual team and ask them and say, Hey, is it you? Because they're all using a different tool. So each team has to go and look at their tool and look at their data and be able to come back and say, no, it's definitely not us. Because they just don't have that that one place they can go to to look at the total environment and be yeah like, where is the problem? right where is the yeah who's Who's being impacted versus who's causing the problem? And so that's just a very a common ah you know frustration for a lot of organizations you know today. and And that's something that we pride ourselves on being able to help them solve. And I also like how you differentiate it between not just being a single pane of glass. That's okay. Like everybody has been talking about single pane of glass, but the first pane of glass, like, yeah, let's start here. And then obviously we can, we can go to get other dashboards. We can go and look at other tools and then figure it out. But yeah, this is where we start. I think that's, that's, that's really important that everybody's looking or starting at the same point and then they are diverging and then figuring out what's wrong. Interesting. Yeah.
00:19:48
Speaker
Yeah. So I wanted to ask you about like, all of this is great, but now, especially over the last two, three years as AI adoption has increased, or at least the the claims about all enterprises using AI models in some way or other has have increased. How does that impact this ecosystem? Right? Like, do do we need a complete rethink of how we think about observability and telemetry, or should we just put like an MCP server in the front and and that's it? Like, and we don't need to make any other changes.
00:20:17
Speaker
Yeah, we're definitely seeing a lot of changes. I mean, like every industry, right, ai is is very disruptive for for us. What we've, you know, realized and has become very apparent is that more than ever, observability now is important. um um There is...
00:20:33
Speaker
you know so much code being written right new you know services being deployed right and the the pace of development has just accelerated and the pressures on teams to deliver faster and faster just keeps it going um and you know observability is that tool to help you be able to deliver faster right like when you're shipping software very quickly right you're inevitably going to ship bugs right like that's just how it is right if you want to move fast you're going to break things and that's okay right like that's the trade-off right You can swap that, you know, speed for, um you know, consistency and and and accuracy. um But at the moment, you know, the world is focused on speed, like deliver faster. And so observability is that tool to help you do that safely, right? So one, you need to have that observability. So when you, when there are problems, you can detect them very quickly. um But most importantly, when things do go wrong, you've got the data to understand what went wrong and why and what you need to do to go and fix it. And so what we're seeing certainly with the introduction of AI tooling and and the amount of software that's being and deployed is that yeah know you just need more observability, right? Especially when you're you're leveraging AI tooling to write the software, right? Like you've got even less understanding of of how these systems actually work um you know than we already had. and it was already challenging enough with microservices to understand you know, the system actually works. But now you've got, you know, these black boxes of code that are just being shipped out. You don't even know what they're doing. And so the observability is that tool to help kind of turn those black boxes into glass boxes, right, so that we can look inside them and see what are they actually doing. And so, you know, we're really kind of seeing this change, right, where, you know, the AI is it is, it is a change, but it's it's more similar to like the move from like building monoliths to building microservices. We just see AI now as more of a, it's a different model
00:22:14
Speaker
paradigm for how you build software, right? This agentic model, right? It's a different architecture. It's a different way of building software, but it's still software, right? um So there are some, some unique challenges and some differences for how you want to observe that, right? um A big one now that, you know, and a lot of our customers are asked for is around, you understanding, understanding, the specific AI things, right? So understanding, you know, yeah is are are these services hallucinating, right? Like how much tokens are they actually using, right? Like how much they cost to run, right? yeah Because right now there's just a big push around like, let's just, you know, don't worry about the cost, let's just go into it. But that is that is quickly coming to an end where, you know, you want to actually understand like,
00:22:51
Speaker
is is it valuable right like yes we're going spending a lot of money on burning tokens um you know but is it delivering the value to our users that we actually expect it to um to actually do right so so there is some some changes in some of some of the data that you want to go and collect and understand um and so you know we've got some tools you know that we announced a couple of weeks ago at our grafana conference around you know just being able to understand like so what we call ai observability right so it's it's It's typical, you know like other observability, like, you know, application observability, but adds in those extra kind of data sets. Then the other side of AI is is how do we use that in actually in observability, right? So we have, how do we observe AI? And then how do we use AI to to do observability? That's also changing very quickly and very rapidly. And, and you know, if you'd asked me, you know, 18 months ago, i would be like, nah, there's not a lot of utility in it. But, you know that has certainly changed, you know, I think,
00:23:44
Speaker
towards the end of 2024 with some of the new models and the shift towards this agentic model. um We have seen a lot of utility in what the AI tools can do in observability. um We've certainly seen this acutely with Grafana. It was actually amazing. We had ah ah ah a couple of engineers you know who worked on a hackathon. So like every quarter, we do ah an internal hackathon where we let the engineering team just go and build whatever they want for a week and then come and share it. And so we had a team that built what is now our Grafana assistant, um you know which is the the kind of... heaven Chat integration, excuse me. Yeah. Okay. So it's the, um, the chat integration we haven't, so they just built it in a week, um, and put it together, like integrating, um, uh, you know, some of the anthropic models, you know, into Grafana and, and they demoed it and we were like, this is incredible. Like yeah how does this work so well? and they were like, we don't know. We just know that it does. Um, But, um, and that was really exciting. Like it was such a great, um, like experience. of This is amazing. We're like, Hey, we've got our next conference, we're finally kind conference in three months time. You're going to be on stage announcing this product.
00:24:47
Speaker
Go, you'd like just build it. Right. So we, you know, you funded them with a, with a team to go and get, get it done. Um, um But as we kind of explored and understood, we we came to understand why it had worked so well because we we were worried as a company. We're like, hey, we don't have a lot of AI capability. You know, like, what do we do? We're kind of falling behind. And then suddenly we just leapfrogged ahead. And and what we discovered is it was down to the open source nature of of what we do at Grafana Labs, right, where we're we've spent, you know, more than a decade building this great open source ecosystem.
00:25:16
Speaker
You know, we've got tens of millions of users around the world who use and love our technology, but also they write about it, right? So they write tutorials, they write blog posts, right? They've got, you know, um public Git repos with their dashboards or, you know, with, you know, material about how to use Grafana, you know, our our code is open source, right? Like, and there' this big ecosystem of all of this content that exists on the internet. And that's the content that the, you know, um foundation models are trained on, right? So out of the box,
00:25:41
Speaker
Whether you're talking to to Gemini or OpenAI or Anthropic models, um they know what our users are trying to do They understand the problem, but they also know how to use our technology to solve that problem. So we've had this really great advantage where we haven't had to go and spend a lot of money on training models and and hiring very expensive data scientists um you know We can just use the foundation models and focus more around you know a nice integrated solution that's in Grafina. Focus on you know adding you know the the the prompt engineering, right adding skills, you know just to make sure that we're guiding you know the LLMs down the right path. And so that's led to us having this really great capability um that delivers a lot of utility. right You can go and ask your assistant to build you a dashboard and it will just build a dashboard for you and it'll make it look really nice. You can go and ask it to run an investigation, say, hey, I've got a, you know, a service that's not working. Tell me why. And it can go and query your metrics and it can query your logs and query your traces and and tell you, you know, you go through that same path ah that your SREs would to understand what the problem is um and and give you a suggestion. And then, you know. we can then let that integrate, right? We say, hey, great, go and connect that to some other MCP servers that have other data or, you know, internally, we're like, hey, we'll connect it to GitHub. So you can just go and look at the code and and go and tell us what lines of code we should go and fix. And so it's really powerful for for what it can do. And that's helping a lot with, you the fact that we've got all these volumes of data and LLMs are great at processing that data, right? There is some challenges with it, right? Because, know,
00:27:06
Speaker
observability data is mostly unstructured data and it doesn't make a lot of sense to LLMs all the time. And so a lot of the work we're we're doing to support our customers is around turning that data into something that makes a little bit more sense. So we have something in our platform called our knowledge graph where we analyze the data and we look at you know the label sets and the metric names, et cetera. And open telemetry helps a lot with this with you know understanding the naming, you know the conventions to know what the data represents. And so we can build essentially a knowledge graph of all the different entities that are represented and how they relate to each other. And then when you feed that into the LLM, then suddenly it becomes really powerful. It knows what data to go and query when um and can give you some really, really um amazing results.
00:27:46
Speaker
That's awesome. So, okay, you said like since Griffana is was open source, all these models were kind of trained on it. um Do you see that the code that's being generated through cloud code or codex or any of these models includes observability by default or built in rather than like added on later? Or do you see like, ah okay, at least what I've seen in the ecosystem is security vendors use ah hooks that cursor provides or to, to, to make sure that these are steps that are always run. So they run vulnerability scans on all the code that is being generated to make sure that nothing nothing makes it to production without a security scan. So do you have to do that for observability that, hey, let's make sure that this code is ready to be observed or the right metrics are being exposed? Or since the models are already trained, they know on all of these open source projects, they know that these are the best practices. This is what you should be doing anyways.
00:28:36
Speaker
Yeah. So I think both of those are true, right? So you still have to ask them to instrument the application. Um, but if you ask them to do it, they know how to, um, right. So, you know, they, they understand, especially open telemetry, right. So well understood, well documented, you know, set of standards and, you know, there's libraries for every code base, um, for some, you know, for some, uh, software languages, right. You can use auto instrumentation, right. Where you don't actually need to do too much. It's kind of just like loaded into the, uh,
00:29:02
Speaker
ah the runtime and and it will, you know, gather the data for you. um But um it's still a decision you need to make. Like what is the correct data that I need to go and collect? um You know, if you're just doing um ah distributed traces are really powerful, you know, you can just like, i'm just going to grab all the spans and understand especially if you're building lots of small little microservices, right? Like you just kind of want to collect kind of, you know, uh, you did the service respond correctly? You know, what was the latency, et cetera. And then you just send that as a, uh, you know, distributed trace, and then we can generate metrics out of that for you and and give you some really good insights. as well as having the the raw traces. But yeah, definitely, you know, if you go and ask, you know, your code code or, know, codex, whatever it is and say, hey, can you instrument this? It will be able to do that for you because it is, ah ah you know, having the the open standards, you know, just makes that a well um a well-und understood problem. Yeah. And there's so much code out there that that ah that these models have been trained on that are already, you know, instrumented, right? So they know what they should be doing.
00:29:58
Speaker
Okay. And the chatbot that you described, right?
AI Integration in Observability
00:30:01
Speaker
Like something that the team came up with in a hackathon, like, is that something that's available only to the Grafana Cloud Platform users? Or this is a feature that I haven't checked Grafana Reads Notes for a while. Yeah. Yeah.
00:30:14
Speaker
yeah ah A bit of both. So it's certainly it's part of our Grafana Cloud platform, right? for ah For our customers to go and use that. We did announce last month at our GrafanaCon where you can use it um in your open source Grafana running in your own environment. You still need to connect it to our Grafana Cloud because there's a lot of like backend services that we have to run. But we do have a very generous free tier where you can get a lot of value out of it um and you can use it to um you know help using it. But you can use that locally in your own environment.
00:30:43
Speaker
yeah grafana that you're running in your in yeah in your home or you know to monitor. That's awesome. Yeah. yeah Whatever you're doing. Yeah. and and And the models, like, do you also do like BYOLM, I guess, or do you only like allow users to connect to like, hey, give us your OpenAI keys or give us your Vertex AI keys and all that?
00:31:03
Speaker
Yeah. So right now we we run all the service ourselves. So you just will connect to Grafana Cloud and and we'll deal with with the models that will get used. And we do get a lot of requests, you know, demand for like, oh, i want to be able to bring my own model. Yeah. It's challenging, right? Cause there's a lot of work that goes into, um, you know the, the prompt engineering and making, you know, optimizing it for the model that we're using. And, yeah know, we make and decisions around which model is the right model to use for the specific tasks that we're doing. So there's a lot of work that goes into it. So right now, um, you know, we just kind of control all of that and you just treat it as a, you know, uh, you know, you're asking a question and we'll, you know, we'll give you the answers. You don't need to know the specifics of it just because things are evolving so quickly, but we want to you know maintain you know the quality of you know what the actual feature can do. And so the best way to do that is that bus just to control exactly what models it's using and how they're using them. Yeah.
00:31:55
Speaker
Gotcha. Makes sense. And you you you spoke about how agents or at least chatbots and things like that can help you ah troubleshoot, right? Like, hey, go go look at different pieces of code and and figure it out. Do you think that's the future? Like, since you are like talking to a lot of customers, is that the future that they see yeah and that... one and that they want to implement like, hey, we'll just have a bunch of agents running around acting as SREs. They are the first call, right? Like L1 support and they'll do all the troubleshooting. If they need help, they'll bring in humans or they still feel that, you building interfaces that are easy to consume to you and humans are is is is also important.
00:32:35
Speaker
Um, we're definitely seeing a change, right. It's, it's definitely, uh, you something we're clearly aware of, right. Like ah as an organization, and you know, that's really started and focused on the dashboarding, right. Like the, the user interface that people go through, we're seeing less and less dashboards are important, right? Like the user interface, like, like everything now is the the chat window, right? So we are seeing that shift where more and more people, yeah know, they might just be in code code. And so they might just say, Hey, you know,
00:33:00
Speaker
what's happening in the service and they'll just go and ask it. And we have some um some tools to help make that easier. So we have a project called GCX, which is kind of like a command line tool to be able to talk to yeah all of our systems. So you can go and query data or you can go and just talk directly to the system or whatever is that you need to go and do um to make it easy for you know your coding agents be how to be able to and do that on your behalf. um But also we see you know within the platform itself, yeah more and more people are relying on the assistant to...
00:33:27
Speaker
ask questions, right? So rather than building a dashboard, you just go into this and and say, Hey, I want to know about this, you know, you know service, right. Or want to know, you know, what, what is just asking a question and will generate a dashboard for you. It'll just generate ah you know some insights for you from the data. um So that's definitely, definitely happening. And obviously organizations that we talk to, you know,
00:33:48
Speaker
you're trying to understand, you know what are their priorities? You know, generally they're all aligned around reduced costs, improve developer productivity, uh, build more reliable systems, right? Like that's pretty much what every engineering organization is trying to do. Um, and you know they see the the AI tools as as a way to help with that, right? it certainly helps with improving the productivity. Um, and if used right, can help with costs. Uh, if used wrongly can just create more costs. Um, but it's about finding that balance. And so we're seeing, you know, how that changes over time. Um, you know Certainly the products that we have in this space are really designed around these things, right? Reduce costs, improve productivity and improve reliability of systems. And that's when we when we look at the products and we look at what the teams are launching and what their features, like that is the lens, right? Like we don't want to release products that don't help with those things, right? we you know We don't want them to be cost prohibitive, right? It doesn't make sense, right? we've we've had we you know we've certainly experimented with some ideas and things we were like yeah this is amazing but then when we look at it like this is so expensive um it's not practical for us to actually deliver that as a product right so we we work on improving the the efficiencies and the you know how it operates to to just align those things um but we are so certainly seeing that trend and and it'll be interesting to see how um
00:35:00
Speaker
you know how this evolves right with you know uh you know the models and and you know how token pricing changes over time and how things uh evolve but right now we're certainly seeing huge advantages to to leveraging ai within observability um just for speed like even just you know the speed at which we can um you know find that that mean time to to resolution for systems right we had a great experience with a uh a customer like a prospect that we were talking to and we were doing a a proof of value with them we were coming in and kind of like set up a kind of well-scoped kind of environment in their system of like, hey, this is how it'll work. And while we were talking about it, they had a major outage and we just live used the assistant to solve it for them and it solved it within five minutes. And they were just like, I mean, this is amazing. This would have taken us an hour to go and solve this problem. And it was just able to do it. It was a real life example. It was such an amazing moment where we're yeah, it's not just demo where I like the there's a lot of utility in in the AI capabilities that we have today. it was just so nice to be able to see that. And so, know, we have now a product we call Grafana Assisting Investigations, right? And where the idea is that you would just tie that in with your alerts, right? And so when you see, you know, if you're using SLOs, you say, we're seeing SLO breach, right? Where a service is not, you know, performing as we expected and it's leading to customer impact.
00:36:20
Speaker
So, you know, you're going to but go and page your your on-call engineer, but at the same time, we'll also just go and kick off an investigation. So, by the time your on-call engineer gets in front of a computer, they should already see a summary, you know, hopefully for what the problem is, right? Or at least, um you know, a good idea of of where to go next, right? if if we If the AI couldn't go and find the root cause for them.
00:36:41
Speaker
So that's what they're like, a usual like 10 step process to do, to troubleshoot something. Maybe by the time the engineer logs in, it has run through step seven and like, yup, these are the results. Now you can just focus on so giving them a head and start for sure. Exactly. Yeah. And then they are, I can't solve every problem. Right. Um, yet, um, i yeah yeah but you know, there's huge amount of value in it being able to solve, you know, 60% of the problem, you know, 70% of the problems for you before you, uh, before you even start looking at things. Right. Um, And so it's really important. And we we still, like we said, we are seeing this shift and we see, um you know, less need for the dashboards. We still do see a lot of use um for the, you know, the dashboarding and for the the visualizations for sure. Because one of the things that we try and do a lot with the AI tools that we're building is making sure there's like an audit trail. Like how did you come to that conclusion? Right. Sometimes it makes mistakes. It gets it wrong. But even when it gets it right, like you often want to go back and say like, well, how did you work that out? How did you know? um and so it's great to be able to have,
00:37:35
Speaker
we keep that that record of like what steps did it go what did the the agents you know what data did they query what queries did they run what responses did they get what conclusions did they make from those responses and often know the way that you know the ai can explain you know what it did is just with a graph and say hey i went and looked at this service i ran this query i saw this graph i saw a spike here um this led me to believe this and then i went and created some more data etc and so you know we still see the the value of those you know the the The observability tooling that we have, right? So the telemetry data we've stored, but also the the way we can present that and and ah you know use Visualize using graphs to help you understand the data it is still just as important as it always has been.
00:38:15
Speaker
Okay, no that's awesome. So, like I know we have been talking about AI a lot. I do have one more question, though. um like You can choose to answer either one of them. I want to understand how engineers at Rafana Labs use AI or how you personally use AI to improve your productivity. like I want to get into some specifics and some things that, I don't know, maybe even I can get started with tomorrow.
00:38:38
Speaker
Yeah, definitely. I mean, yeah, we' we're definitely, as an organization, we encourage everybody in the organization to be using AI, you know, where they can to help accelerate the work that they're doing. and That's certainly true for for our development teams. um I certainly use AI a lot myself, right, just to, you know, accelerate the work that I'm doing. And and ah and it's great.
00:38:57
Speaker
and It's a great tool. um We have actually a great internal tool we call our GTM assistant, which is you know, great for kind of, you know even our go-to-market teams, right, to be able to just ask it questions about our products or about our customers or, you know, what information do they need to know? How do I go and pitch this to this style of user, right? And so there's a lot of great tools that we're building internally on, know, top of AI agents, which is really exciting to to see. and And that's, you know, also what's feeding, know, a lot of the changes we have in observability, right? Because we're we're building a lot of AI, you know, ourselves. And so we're like, how do we understand how this is working, how this is performing? So we have to go and build you tools for us to understand it and then obviously our problems aren't different to anybody else's and so we just you know turn it into a product and share it with the rest of the world um but yeah we certainly our engineering teams we see um i think just about all of our engineers use um you know ai uh you know coding assistants to to get their job done and what's what we're seeing more and more is like the role of an engineer is changing now where more and more they're becoming the engineers are becoming more like product managers, right? Where they're just going to understand the user problem. What is the challenge that the user is facing? And then come up with an idea to how to fix that. And then just tell the the yeah the agent to go and build it for them. um And that's something that's working really well for us at Final Labs. We've always been very kind of engineering led, right? So the engineering teams have always been
00:40:15
Speaker
you know, focused on like that they own the product roadmap. We have product managers and we have a great product team that help facilitate that. They go and, you know, talk to the customers and understand the customer problems and see what's happening in the industry and come and share that and facilitate the conversations with engineers to help them come up with the, you what the roadmap should be. But ultimately the engineers that decide what we're going to go and build, um, you know, for a lot of our teams. Um, and so they've always had good product instincts. And so now we're seeing them, you know, be able to be very effective with the use of AI tools because, They're able to take their understanding of the the problem. I mean, it helps that we're, you know, a bunch of engineers building solutions for engineers. So they're able to kind of understand what the problem set is and just be able to yeah to go build your solutions very, very quickly. And it's great be able to see the the acceleration. We still obviously have to have people involved, um right, just to you know ensure we have consistency around our software, just to ensure that um it's maintainable. And it aligns with how we build software and and and so that we can support it over time and keep iterating on it and improving it. But we do see a lot of um you know benefits from the of AI tools for sure.
00:41:18
Speaker
And like ah has the expectation changed? like As you said, right like PMs, you still have a product team. other Have the expectations on them changed? Again, I'm i'm being like biased here. I'm a product manager. So ah are you expecting PMs to give you prototypes or still it's okay to give you a Jira ticket with detailed requirements?
Customer-Centric Product Management at Grafana
00:41:39
Speaker
like where where Where is the the line right now?
00:41:42
Speaker
Um, I mean, if they can, sure. Um, you know, for the little things we're building, it's, it's more around like just having those discussions, um, and, you know, understanding, you know, what is the problem? big A big effort we focus on is, um, and and the product managers is great at helping with is understanding what the problem is first, right? Before we start designing the solution. Um, you know, we,
00:42:03
Speaker
it's It's a challenge we have. you know you You don't want to let engineers unfettered to just go and talk directly to customers on their own because they'll always just be like the customer will say, hey, I need this solution. And the engineer will say yeah, I'll go and build that for you. And then come back, give it to the customer. The customer's like, this actually doesn't solve any of my problems. And it's like, what? Okay. So it's like really focusing on, you know, because engineers, they want to, they just want to solve problems. Right. But yet you have to be a little bit disciplined to say, like, let's not talk about the solutions until we've talked about the problem first, um, and understand what the problem is, and then we'll decide what the best solution is based on that. Um, but our product, you know, teams, um, you their, their job is really that, that, that really important piece of going out and talking to the customers. And being able to ask the right questions to understand what is the actual problem that you're facing um and and really understand the problem um um and then be able to come back and and share that with our our internal teams so they can go on build solutions. Yeah.
00:42:54
Speaker
Okay, nice. ah One of the things we see a lot is, especially us growing as an organization and especially like the open source community, there's so much feedback that comes in and the AI tools are great at being able to look at all of the data and then be able to summarize it and find trends and be like, hey, we're seeing consistently this is a an area that people you know would like to see some you know changes in or or new features in. So AI is very great at kind of processing all of that data to give us some new insights that previously would have been very difficult for us to be able to to go and see.
00:43:24
Speaker
Gotcha. No, that's definitely interesting. Right? Like sounds like, uh, yeah, a lot of AI is being used for the right reasons at Grafana labs. And yeah, I'm sure people who are just getting started or experimenting with it would would have some nuggets to pull full from this conversation.
00:43:37
Speaker
Um, I wanted to go back to like your experience with like the early, early Grafana days, right? Like having an open source project is great. I know you said, um, one of your co-founders started it, uh, but like if somebody is thinking about starting a new open source project today, how do they scale? Like how do they become the next refund? How do they become the next open claw? Like but some of the best practices or what guidance would you have for, ah for somebody just starting?
00:44:02
Speaker
Yeah, it's certainly challenging. Uh, I don't know. and I don't think there was like a it was very strategic about how we did it. We kind of just did what felt right. Um, and so I think a lot of it is really focusing on, um,
00:44:17
Speaker
it's It's finding that balance, right? Like you want to focus on the community and you want to build solutions that solve their problems. um And then at the same time, you don't want to be too prescriptive, um right? Because certainly with open source, like one of the the benefits and the advantages and why people love it so much is it's their ability to...
00:44:36
Speaker
use it how they want, right? And adapt it yeah for their own specific environment. And so it's about finding that balance of like, you want to deliver utility, um but also you want to give a lot of flexibility for how people use it. Right. Cause that just then creates ah much larger ecosystem and a bigger pool of people that can come and contribute. so that's one of the things that we really focused, you know, certainly with Grafana, right. Was really focusing on, you know um how do we make a tool that solves a problem really well, but also can be used to solve lots of different problems, right? And not be too prescriptive. And we still focus on that today for our open source, right? We consider it our open source to be a set of tools that you can use to solve thousands of different problems in many, many different ways. And then our products are more solutions, right? um And that fits really well with...
00:45:22
Speaker
our open source community where the open source, they want the the the tools because they want to be able to tweak it and customize it and use it how they want. um And then they don't actually want opinionated solutions, right? For their, you know, if they're using open source, whereas those that are are looking at our commercial, you know, solutions, they want opinionated. They want it just work out of the box. And so that's ah a great differentiator that we have. And that's, you know, certainly for ah for anyone building an open source, you're thinking about open source, obviously,
00:45:46
Speaker
You, you know, if you're going out to build it, you know, typically it's because you, you know, you want to you know make money off it eventually somehow. yeah Um, so for us, we find that a good way of, of having a clear delineation between what goes in our open source versus what goes in that commercial and and feel comfortable being able to do that. Right. We don't want to ship, um, you know demoware where it's like, yeah actually all the cool stuff is just in the commercial offering. like We want to make the open source really good for what it does. But also, we're still a business and we need to have ah a way to be able to go and commercialize that over time. And for us, that has been the build tools for open source and and give them flexibility and choice to be able to go and adapt it and use it how they want. um And then for our commercial, it's it's more of that ah solution-based um approach.
Balancing Open Source and Commercial Offerings
00:46:26
Speaker
Interesting. No, I think that that definitely answered then even the next question I had, like, how do you generate revenue off of an open source project? So I think that that delineation is super helpful. Give the flexibility. But then if somebody is looking for that opinionated opinionated solution, this this is where you pay, I guess. yeah um Yeah. And you've got to like, it's it's about delivering the value, right. That people want. And certainly like, um you know, for, for us, like, you know, open source is still so important to the business. Um, you know, we have, you know, tens of millions of users are our open source and we know that to be a successful business, we only need to you know monetize a very small percentage of that. Um, yeah right. Like don't need to try and get every single one of those users paying us money. Um, uh, we want to focus more on building a great ecosystem, you know, getting people, you know, loving, um, you know the products that we build and the, and, you the technology that we have. And, you know we, we talk about like certainly for Grafana, right. We focus on like the ubiquity of Grafana. We just want it to be easy to use for everybody everywhere and everyone to have had experience with it. Right. And had had a good experience with it. Um, And that just helps ah you know us as an organization, right? When they're coming to choosing observability vendor and they see Grafana as an option, like, oh yeah, I know that tool. I had a great experience with that. That's a great option. We should go and look at that. um So that's that's certainly our approach. Yeah. You have champions distributed across the organization that can help you. That's awesome. Yeah. Yeah. It's great. It's a great way to get into organizations, right? Because, um, uh, especially in the big enterprise, right. Where people don't need to ask for permission. They don't need to talk to procurement to be able to just go and download some open source and start using it and get value out of it. It's kind of like, you know, it's like that gateway drug, right? Like, yo, you get a taste of it. Uh, and then you like it. Um, and so, yeah, we see that a lot where like, it just, it brings down the barriers of entry, right? People can get access to the technology that get a lot of value out of it, uh, before they even need to go and talk to anybody. Yeah.
AI and Future of Observability Solutions
00:48:09
Speaker
Gotcha. No, that's awesome. Okay. Last couple of questions. um What's next for Grafana Labs? I know you said you had a the annual conference a couple of months back. I'm sure we'll link to ah YouTube channels where you have those playlists already created so people can um can go watch those. But what's something that the Kubernetes community should be looking out for or keeping an eye out for from Grafana?
00:48:33
Speaker
Yeah, I mean, we're definitely going to continue just doing what we've been doing, right? Focus on building great observability solutions, um you know, and just kind of completing, um you know, what's you know yeah your observance observability you know needs and and solving for those, you know big focus right now is obviously AI, both from ah how do we leverage AI in observability, but also how do we help our customers?
00:48:54
Speaker
understand all of these AI agents that they're going and building. um So that's ah a big area. And that is, you know, obviously everyone's, you know kind of getting involved in that and and building agents. Lots of, lots of organizations are seeing that as a path of of how they can build, um you know, you know more solutions. I mean, for me, I think, I feel like the agents, it's it's no different to how we've always, you know, looked at trying to automate processes. It's just that now, you know, your sales team can go on, you know, ask, you know, call code to build on an agent, right. And they can you know get value out of it. And so we're just seeing more and more software getting, getting written. So we want to just help manage that. So that's both with, the capability, right? So to understand, you know, what these things are doing, but also help manage the growing costs, right, of of this huge volumes of data now that they're generating so that we have to understand what's happening. And so that's a big area of focus focus for us. um Kind of next for us is kind of, you know, we're exploring more into ah kind of the BI space. We know certainly in open source, Grafana is you know, loved and used in a lot of BI i use cases. There's a lot of BI data. Now you want to kind of bring in and understand, right? you You know, we're seeing even just observability, right? What that means is just expanding, right? We're observing our applications in our infrastructure, but also, you know, more and more we want to observe our business processes, right? Understand what's happening inside of our organization. And so building more kind of solutions in that space is is an area that we're exploring just to yeah We've seen it be very successful in open source. And so like we want to be able to help commercial customers. And and a big part of why we're seeing that work really well is LLMs are really great at writing SQL. So um you know when you've got a whole bunch of data stored in you know your data lakes or you know big solutions, we want to be able to kind of bring that in and and you know be that be part of your understanding of what's happening in your environment. So that's a big area of focus for us as well.
00:50:34
Speaker
Gotcha. Okay. Awesome. And last one last question, where can people find you on socials? Do you have a medium or a blog post that you own and then publish on like, how can people learn more from you?
00:50:46
Speaker
Me personally, I mean, LinkedIn is about it. Um, uh, I'll post on there every now and then. um you know, I have some other you know blog posts that we put out just on the Gafana blog. Um, there are people far smarter than me who write some great content, um, um, that I can, I can read and I can share with you. and yeah, I'll, On our blog, um if you go to the Grafana.com website, um there's a bunch of content there you know to help understand what what we're doing, written by you know great people internally. We also have external people write for us, you know whether it's customers, you're part of our open source community know who share what they're doing. um But if you go to the Grafana.com website, from there is probably a great place to go, whether it's our YouTube channel or um ah you know our blog posts or other articles that we publish around the place.
00:51:31
Speaker
Oh, that's awesome. Thank you so much. I think this has been a great conversation, Anthony. Thank you so much for your time today. Thank you very much. I've ah really enjoyed it and I appreciate you having me on. That was a great interview. Hope you guys liked it as well. We do have our next set of guests lined up. So I'm excited to bring you a cool set of episodes over the summer. And we would appreciate if you like, share and subscribe the podcast, ah to the podcast regardless of whether you're listening to the audio version or watching us on YouTube. um And with that pitch, it brings us to the end of another episode. I'm Bhavin. And thank you for listening to the Kubernetes Bytes podcast.
00:52:11
Speaker
Thank you for listening to the Kubernetes Bytes podcast.