#125 October 13, 2020

Okteto, with Ramiro Berrelleza

Hosts: Craig Box, Adam Glick

Ramiro Berrelleza is CEO and co-founder of Okteto, a company making developer tools which simplify development on Kubernetes. He joins Adam and Craig to discuss how the open source project and company came about, going through Y Combinator, and the best filling for a Mission burrito.

Do you have something cool to share? Some questions? Let us know:

Chatter of the week


News of the week

ADAM GLICK: Hi, and welcome to the Kubernetes Podcast from Google. I'm Adam Glick.

CRAIG BOX: And I'm Craig Box.


Happy birthday, my friend.

ADAM GLICK: [LAUGHS] Oh, thank you very much.

CRAIG BOX: How did you celebrate the event?

ADAM GLICK: Well, it was a wonderful day here. My wife was very nice and made my favorite breakfast for me. I got a chance to play with our little one and just be a little bit lazy around the house, sleep in, and enjoy the time that you don't often find.

CRAIG BOX: What is your favorite breakfast?

ADAM GLICK: So I was going for some hash browns and some pancakes. And it turns out-- I tried to make the hash browns, and that didn't work out so well. Turns out making hash browns is actually hard. If you just take and shred a bunch of potatoes and then put it into a fry pan and fry it up, what you end up with is this kind of dark brown, kind of starchy mass. Which doesn't taste that bad, in fairness, but was not the hash browns I'd aimed for. So next time, I'm going to have to look up a recipe on that one. The pancakes that my wife did were far better.

CRAIG BOX: What's the glue ingredient in the hash brown? Do you need some egg or something to hold it all together?

ADAM GLICK: Normally, you want them to be separate and fluffy. So you actually want to rinse the starch off of them so that they don't glom together. So you rinse them, and then you want to dry them. I, of course, looked all this up after I had my hash brown accident rather than beforehand, which would've been much more useful. But next time, they'll be better.

CRAIG BOX: I did try to make some corn fritters recently, and I underestimated the amount of flour in them. So they ended up kind of gloopy, but tolerable. That's the best that describes it. They were tolerable.



ADAM GLICK: Welcome to the podcast CI/CD loop. Of course, since we're both humans and living in meatspace, we have a certain level of non-automatable activities. This is where we'd love your help. Craig and I are always looking for how we can make this podcast more valuable to all of you who are listening. For those of you who have met us, we often ask what we can do to make the podcast better.

CRAIG BOX: In that spirit, we're introducing the first Kubernetes Podcast from Google audience survey. It's a short set of questions to help us get to know you, your listening habits, and your feedback on the show so that we can continuously improve. We think it will take less than three minutes to fill it out, and your answers will help us make sure the show remains the best podcast for the Cloud Native community.

ADAM GLICK: Please go to kubernetespodcast.com/survey, or click the link in the show notes. And tell us what you think. It really does have an impact on what we do.

CRAIG BOX: Let's get to the news.


ADAM GLICK: Congratulations to the Rook team as the latest project to graduate from the CNCF. Rook is an orchestration system that connects multiple storage options like Ceph, NFS, CockroachDB, and more, to Kubernetes. This brings the count of graduated projects to 13, with Rook being the first graduate that focuses on storage. Rook came to the CNCF in 2018 from Upbound, and you can learn all about it in episode 36 with Upbound's Jared Watts.

CRAIG BOX: Just in time for EnvoyCon this week, project founder Matt Klein has announced that WebAssembly, or Wasm, support has landed upstream. Wasm extensibility had been developed in the long running fork by the Istio team at Google Cloud. And this is an important step in making it available by default to all Envoy users. This comes just after the release of Envoy 1.16, and documentation is expected to land in time for the release of Envoy 1.17, due in a few months.

ADAM GLICK: Helm charts are on the move, at least if they're stored in Helm Hub. Helm has announced that they are moving their chart repo into the CNCF's artifact hub, a more general open source storage project. Artifact Hub itself is a sandbox-level CNCF project, and shows another example of how our community is growing and building a broader ecosystem using its own open source tools. Any charts you had or used in Helm Hub should be available in Artifact Hub, and the old repository now redirects to the new one.

CRAIG BOX: DigitalOcean has introduced a DigitalOcean App Platform, a new platform as a service. DOAP, which they are hoping you call "Dope," is built on their Kubernetes service and uses gVisor, Cilium, Buildpacks, Kaniko, Istio, Envoy, Prometheus, and Fluent Bit, to name but a few projects in the stack. The service will deploy from code with bring your own container on the roadmap. It is now live in three regions.

ADAM GLICK: IBM has announced that they are splitting their managed infrastructure business out into a new separate company. What remains of IBM will focus on high growth hybrid cloud and AI, while their managed infrastructure business-- or what some might have called their IT outsourcing business, if we use terms from two decades ago-- will spin out into a yet unnamed separate company. The Kubernetes and Istio work and Red Hat technologies that our community is familiar with will stay a part of the new IBM.

CRAIG BOX: The Kubernetes steering committee election has concluded for another year. Congratulations to Davanum Srinivas from VMware for his re-election, and to the two new members of the committee, Bob Killen of Google, latterly of the University of Michigan, and Jordan Liggitt, also of Google, latterly of episode 34.

ADAM GLICK: If you have the steering bug, the OpenTelemetry project has just announced an upcoming election. Four community members will be elected to their governance committee, which is gradually replacing the bootstrap committee put in place when the project was founded.

CRAIG BOX: Le Van Nghia of CyberAgent has announced the open source release of PipeCD, a declarative continuous delivery tool for Kubernetes, serverless, and VMs. The goal of the project is to create one CD tool for all the different ways applications could be deployed. The initial release focuses on visibility, automation, security, and multi-tenancy/multi-provider.

ADAM GLICK: Anchore, with an E on the end, has announced the aptly named Anchore Toolbox, a collection of open source DevSecOps tools. The Toolbox is just starting to get its collection together and starts with two tools. The first is Syft, with a Y, which scans containers and source code directories. The second tool is Grype, also with a Y. Grype scans projects and containers for known vulnerabilities. Together, the tools are designed to enable the growing shift left trend in security.

CRAIG BOX: Rancher Labs has announced the release of Rancher 2.5. The latest release provides the ability to install Rancher's agent on any cluster, allowing use of their controlplane with multiple Kubernetes distributions. If you're using EKS, Rancher is looking for you, as they now provide lifecycle management on that platform. Additionally, RKE Government was announced, which is a version of Rancher for governmental agencies, using SELinux as well as providing FIPS-140.2 certified encryption. Finally, bundles were announced as a new way to deploy and manage applications.

ADAM GLICK: Red Hat has announced a price drop for their dedicated open shift offering on AWS, Google Cloud, Azure, and IBM Cloud. The package can now be purchased as a three-year commitment with a cluster fee of $263 per year and worker nodes starting at $667 each. Red Hat also announced an increase in their SLA to 3 and 1/2 nines, or 4 nines if you're using IBM cloud.

CRAIG BOX: Kubernetes continues to fly high in the defense industry with an announcement from the US Air Force this week that Kubernetes flew on a U2 reconnaissance plane. Neither Bono nor the Edge were spotted or harmed.

ADAM GLICK: Finally, Eric Jadi this week released a UI for Kubernetes that is best described as how Old MacDonald would tell his kids about Cloud Native. He's created a management experience for Kubernetes in Minecraft where pigs are pods, cows are replica sets, chickens are services, and so on. What it may lack in efficiency, it more than makes up for in creativity and 8-bit fun.

CRAIG BOX: And that's the news.


CRAIG BOX: Ramiro Berrelleza is the CEO and co-founder of Okteto, with a K, who build tools to improve the developer experience for Cloud Native developers. Before this, he was an architect at Atlassian and an engineer at ElasticBox-- no relation-- and Microsoft Azure. He is originally from Mexico, and currently lives in San Francisco. Welcome to the show, Ramiro.

RAMIRO BERRELLEZA: Hi, Adam. Hi, Craig. Thanks for having me. Very excited to be here.

CRAIG BOX: What got you interested in technology?

RAMIRO BERRELLEZA: I got into technology when I was very little. Video games was kind of the gateway to me. I remember I got a Nintendo back home, and there was this game, Excitebike, that really got me. I remember you could build your own roads. That's when I got this sense of, huh, you can actually create things with tech.

Then after that, video games got to computers. That got to learning how to code, taking one C++ class in my senior year of high school. And then when it came to pick my degree at college, I just went with computer science and I have not been able to get away from tech since then.

CRAIG BOX: Have you at least been able to get away from C++?

RAMIRO BERRELLEZA: Yes! I am a happy Go developer these days. Or, well, used to be. Now I'm more of an email developer, but you know. [CHUCKLES]

ADAM GLICK: [LAUGHING] Indeed. You started out at big companies-- you worked at Hitachi, IBM, Microsoft-- and then you moved into the startup world, where you've been ever since. What inspired the shift from big company to the agile startup world?

RAMIRO BERRELLEZA: For me, it was two things. One was being at the right time and at the right place. Two of my good friends at Microsoft, Alberto and Amadeo, started ElasticBox. They moved to the Bay Area, and then they started reaching out to me, say, hey, come join us. It's a lot of fun. You should try the startup life.

And it was something that I always knew of. There's a lot of mystique around San Francisco, Silicon Valley, startups. Back in Mexico, the tech industry there, startups were not a thing. After my fourth year at Microsoft, "yeah, why not." Let's try this. I'll do it for a couple of years, see if I like it. And then once I moved here, I started working in startups.

Now I'm like, I'm not going to go back to a big company. It's just such a different world. ElasticBox was such a wild, fun ride that made me really want to stick around for more of this.

CRAIG BOX: Tell us about that wild, fun ride, because as employee number 5 with ElasticBox, you may have done quite well out of the fact that the company was acquired two years after you joined. What did that make you think about future startup participation?

RAMIRO BERRELLEZA: That was a lot of fun. I was one of the first engineers there when we joined. It was the founders who had an idea, a vision, and then my now co-founder Pablo and I were pretty much the first two engineers. So it was a lot of work, a lot of highs, a lot of lows. But it really showed me what a small, focused team can really accomplish. I came from big companies where roles were very dedicated. It was like, oh, you are the engineer for this small piece of the whole world. Then you move to a startup and you're like, yeah, it is! No, it is, and it was for me!

It was a huge adjustment. I went from Seattle, having a PC, writing code in .NET, having a Windows phone, moving to San Francisco. And it was like, what's this Mac thing? OK, where's the Windows key? Oh, right.

CRAIG BOX: Yeah, they check you at the border for California, don't they? "Oh, a Windows phone, I don't know you're allowed that."

RAMIRO BERRELLEZA: [LAUGHING] I was able to sneak that one in. So many people made fun of me in San Francisco for having a Windows phone. But I still miss my Nokia Lumia, I got to say. [LAUGHS] It was just this big change where you now own all these things. You have a lot more impact. And you can really help deliver this vision in a much more hands-on way that I really like about startups. And that was what ElasticBox really showed me you can do. And that you can have a lot of fun while doing it with your friends.

ADAM GLICK: You get to wear a lot of hats at a startup as opposed to in a large organization. Most recently, you've moved on and founded Okteto. What inspired you to create your own company?

RAMIRO BERRELLEZA: That's an extension of what happened at ElasticBox. Okteto is a company that I started with two of my closest friends, Pablo and Ramon. The three of us came from ElasticBox as in typical Silicon Valley fashion. You're going to start building your group, your friends, you're going to bring them along.

I worked in a big company at the time. I was at Atlassian. My two friends were at Docker and at Google. And we just kept talking about all these cool things we wanted to do, and all these cool projects we were trying to push through in our companies.

And one day, Ramon and I were having dinner somewhere in the Mission, and it was like, you know what? Why don't we just do our own company? It'll be so much fun. Pablo was living in Spain. So it was like, we can just move to Spain, spend a summer there, start on this company, and then if things go well, we have a job. If not, we'll just find something else-- number 1, working with my friends, and number 2, having all these ideas we wanted to.

CRAIG BOX: Spending a summer in Spain.

RAMIRO BERRELLEZA: Yes. That's the real reason why we have this company. Everything else is just, OK we do have to do something for a living. But I'd need an excuse to go spend a few months, there, which I'm not going to say wasn't nice.

CRAIG BOX: We do like to ask people whether the problem or the company came first, and it does sound like that you got a few people together and said, we like working together. We want to build something. How did you go about identifying which problem to attack?

RAMIRO BERRELLEZA: By elimination and trying a thousand different things. It's a funny story, because when we got together, the first thing that I said was, I'm done with dev tools. I don't want to do this anymore. Let's do something new. And they were like, OK, cool. Yeah. We'll start experimenting with everything from voice-based apps, IoT devices. At some point, we're about to join a VR, LaunchPad Accelerator just to do something. And then one day, one of the mentors there told us, you don't know anything about this space, and you clearly know a lot about dev tools. Just go do that.

And as we were building all these other prototypes, we kept building this tooling to help us build these prototypes. And it's actually where our open source project came from. At some point we were like, OK, let's accept this. What we like to do is dev tools. Let's just go at it.

By then, we had some experience with Kubernetes, and we have seen all these pain points that our friends complained about, that we suffered personally. And then we saw a good chance to say, we can contribute a lot by taking our expertise on dev tools, on Cloud services, and really put it to the task of making it easier for everybody else to take advantage of this wonderful set of technology that is Cloud Native and Kubernetes.

ADAM GLICK: Didn't dev tools pretty much just max itself out when someone created Emacs? And everything since then has mostly just been plugins and other tools.


CRAIG BOX: That's a very Vegemite opinion of you, my friend.

RAMIRO BERRELLEZA: Yeah, especially because I'm a Vim kind of guy, but yeah.

ADAM GLICK: Fair enough. Not to open up that debate, but tell us a little about what Okteto does.

RAMIRO BERRELLEZA: Right. At a very high level, at Okteto we're building tools to make it easier for developers to use Kubernetes as a developer tool. A lot of what we do is hiding Kubernetes. Most developers should see Kubernetes as this compute infrastructure, hat gets you apps running, but then they should be focusing on their own applications. And that's something that early tools in the cognitive space have been more focused on, making it easier to access Kubernetes.

And for us, it's about the next step. It's, now that you know how to run a cluster, now that you have your apps running there, how do you make your team a lot more efficient? How do you make your developers happier by giving them all these tools? Basically, they don't have to edit YAML everyday.

CRAIG BOX: But people love YAML. Why would you take it away from them?

RAMIRO BERRELLEZA: I'm not a big fan. I have strong opinions. I don't like the spaces in my code. I don't like Python for the same reason. [CHUCKING]

ADAM GLICK: It's one of those things that is always going to be changing. Whatever the data form is-- XML, it's JSON, it's YAML-- there's always a new language, and there's always a new file format.

RAMIRO BERRELLEZA: We love reinventing those things. It does get better over time. I think we can all agree that YAML is much, much better than writing XML.

CRAIG BOX: Had you been working pure engineering roles up until this point?

RAMIRO BERRELLEZA: Yeah, this is my first non engineering role. In Atlassian, I was an architect, but before that, I was a principal engineer. So I was writing code all the time. Microsoft was the same thing. ElasticBox was pure coding. So yeah, this is my first adventure as a non-developer. I still code some of the time, but it's no longer my main role.

CRAIG BOX: But when the three of you sit down in your San Francisco restaurant to discuss the company, how do you decide who's going to be the CEO?

RAMIRO BERRELLEZA: I think I just missed that meeting. Then they decided that, oh you're not there? You're going to have to deal with all of this.

CRAIG BOX: Oh, two out of three, sorry.

RAMIRO BERRELLEZA: Yeah, Ramon and Pablo live in Madrid now. So I think they had a secret meeting without me. No, it was a function of who was better suited for different roles. And Pablo's a great CTO, and Ramon had this great vision of products, and is driving our product efforts. And I felt like me, by being in San Francisco-- especially because we were thinking about having to do fundraising, engaging with more customers-- it just made sense.

ADAM GLICK: What's your main job? You mentioned fundraising, and I guess in a startup that's certainly a core part of what you need to do. What does that look like?

RAMIRO BERRELLEZA: That's an interesting question. I don't think I know at this point. I'm trying to figure out what a CEO is supposed to do. A lot of what I do day to day is being the external-facing part of the company. I talk a lot with our users, write content to make people understand better what we do, talk to investors, of course, talk to other companies for partnerships. So a lot of what I do is more, how can we drive our vision on the outside while the rest of the company, my co-founders, our dream team, builds all the features and releases a product. That's how it's split right now. But one thing I like about early stage startups is that it changes so much week by week.

ADAM GLICK: Do you ever miss doing the tech work? Do you ever think, hey I want to get out of email and more into Emacs, or whatever your coding environment of choice is?


CRAIG BOX: Have you learned nothing?


ADAM GLICK: You can go Nano, you can go Vim, you can even go VS Code, whatever you like.

RAMIRO BERRELLEZA: That is true. These days I'm a VS Code fan. Yeah, I do miss it every now and then. I do write some features. I do like writing code. I'm a developer that writes code in their free time. I really enjoy coding, and I have all these hobby apps that I build in my spare time.

One thing that our mentors have been teaching us is how being focused on what the company needs can really help you advance your vision. And sometimes that's not coding. And it's been that we as a company need to learn, need to build. I'm also enjoying this facets of my careers. It keeps me busy, keeps me excited.

CRAIG BOX: Speaking of mentors, Okteto was part of the winter '19 batch at the Y Combinator accelerator. How did that come about?

RAMIRO BERRELLEZA: It's a cool story. A few years ago, Pablo interviewed for Bitnami. He didn't get the job, but he became friends with the founder of Bitnami, Daniel, who's also from Spain. When we were playing around with the idea of Okteto, building prototypes, at some point we talked to him. And his first thing was like, yeah you should join YC. I think you're a good fit. It's going to help you a lot in what you're doing. I think he saw himself in us, a little bit. He's kind of our first mentor in Okteto, and he was really the reason why we applied.

We didn't have a company back then. We had this open source project. We had some ideas of what we wanted to build. And then we just spent a few weeks doing the infamous one-minute video-- which takes hours to make, of course-- fielding all those questions, which was a really cool process. And then we applied, and we were lucky enough to get admitted.

CRAIG BOX: Did that mean that Ramon and Pablo had to move back from Madrid?

RAMIRO BERRELLEZA: Yeah. The three of us moved here for the duration of the batch, which was three months. We ended up renting a house in the Mission, of course. I can't seem to escape the Mission. So we rented this very nice house in Folsom and 24th, which is right where Philz Coffee is, for those of you who know SF. And this is very mythical. It's kind of this cool hacker house. For us, it was a pretty cool experience. And then going to YC, working on the company, working on the product, doing all these different things. And it was a great experience. I really recommend that to everyone. I really enjoyed it.

ADAM GLICK: How did your significant others feel about that move?

RAMIRO BERRELLEZA: Well, my partner lives here in San Francisco, so for her it was not a big deal. But yeah, especially Pablo, Pablo had a small kid at that time. Well, he still has it. It was smaller back then. [CHUCKLES] They made the sacrifice, and I really appreciate how they both made sacrifices. They were flying back and forth all the time, but also their partners-- I'm most thankful that they allowed them the space. And they also made the sacrifice to let us pursue those dreams. It was a big happy family where everybody was helping each other.

CRAIG BOX: At the end of these accelerators, you go through a Demo Day process. What was that like?

RAMIRO BERRELLEZA: Whew, Demo Day was brutal. It's one of those things that I've seen videos of before. I've heard people talk about it, but I never could have imagined what it was like. It was very intense. You have this one-minute pitch where, at that moment, you feel like your entire future depends on that pitch. Now I know that it's not true. But back then it's how it feels.

They're rushing you through like, go!. And they have you talk to hundreds of investors. Everyone wants to get your entire company in 30 seconds. So it's very intense. It's fun. Those three months, they prepare you for this day. And it is significant, because that's when you raise capital that allows you to grow faster, hire more, build more features, all those things. A cool experience, but it was something that I'm glad that I don't have to go through again.

ADAM GLICK: It is an experience that not too many people get to go through and a lot of people want to be able to go through. What's the most valuable thing that you learned that you could share with the people who are listening?

RAMIRO BERRELLEZA: For us and for me personally, the most valuable thing-- and this comes back to me being an engineer for most of my career-- is there's a big difference between building a product and building a company. And the three of us were really good and really experienced building products. We've done that for a long time. But building companies is a whole different thing.

YC is really good at helping you make that transition and learning what it takes. It's not just about your features. It's about the story you're telling. It's about your vision. It's about how you're going to change the world. Some of my batchmates were, oh, we're building an electric plane! We're going to solve global warming. For us, it was more, I want to help developers be more efficient at their job.

We all have our dreams, our visions, but YC is really good at teaching you. It's your dream, and make sure you sell it, and make sure that whatever you're doing, your company is advancing that goal. It sounds obvious maybe in retrospect, but when you're deep in the weeds, writing code, building features, it's not as obvious.

ADAM GLICK: Let's talk about the project itself. People often talk about Kubernetes and about whether it's a developer or more of an operator tool. You have a lot of experience with developer tools. What do you think about the current state of developer tools for Kubernetes?

RAMIRO BERRELLEZA: Right now, it's a very exciting state, because we are in the beginning of Kubernetes. I know that for a lot of us, it seems like a long time. I was lucky enough to be at DockerCon the day that Craig McLuckie showed Kubernetes to the world. I remember that day and I was like, yeah this is really cool. But that's for the ones who are super deep into it. For most people, it is just the beginning.

It's very exciting that you're starting to see how all these tools are being built in place. I think the first few years of Kubernetes was all about making it easy to operate, easy to deploy, easy to run things. And now, we're in a time where there's a lot of people, a lot of companies like ours, trying to solve the next problem, which is now that you have a cluster running, now that you know how to scale your application and keep it running in production, how do you make it easier for developers to use it and to build apps-- because if you look at how we're writing apps, even before this, when containers came to be or the whole DevOps revolution, that changed how we deploy, but that didn't change how we write code. We're still using VS Code-- or Emacs-- locally on a single machine. Everything is running bare while we're building these apps that are supposed to be scaled up to infinity.

I think the new generation of tools are going to take care of this and really make developers expand into Kubernetes and really use it as a dev platform in a way that's easy, in a way that's accessible, because right now, Kubernetes is still very much for experts. And I think that's something that's going to change over time. We saw the same thing with Cloud computing when Heroku came out, when the early HashiCorp tools came out. I think we saw the same transition. And I think it's going to happen. And hopefully we can be a small part of that.

CRAIG BOX: Do you think that the development experience is broken in Kubernetes?

RAMIRO BERRELLEZA: It's a strong word. I like to use it because it's the expression "broken" that I think provokes talk and discussion. I think that today, if you want to build an app for Kubernetes, you have to keep a lot of things on your head. You have to understand everything from your app to the container model to YAML to network policies. It's too much to ask a single developer.

And I think that's where this next generation of tools are going to be helping developers, not having to worry about this. For me, Kubernetes, in the same way that Cloud-- they should be invisible. You should be able to write a Java app, a Go app, whatever it is, without even seeing Kubernetes. For you, it's, that's where my app runs, and I get all these benefits, but I don't have to build with it. I think what Netlify did for the Jamstack is what we need to do for the rest of Kubernetes where it just runs. And you focus, as a developer, on your applications, on your business domain, on making your customers happy, and then Kubernetes just takes care of itself or through tools like what we're building.

ADAM GLICK: It sounds like there's a lot of view on, how do you abstract it away, and keep developers focused on writing the things that they want to write, which really makes Kubernetes more of an operator tool and something that operations is really worried about running. In that aspect, why is it so important for the community to continue to invest in developer experiences?

RAMIRO BERRELLEZA: I think it's important because developers are moving towards more complex applications. And developers should be focusing on that aspect, on the application. If you don't have a good dev experience, then you have all these abstractions leaking, and it's more things you have to think about. It's more things you have to take care about.

One of our early demos of Okteto was this idea that today, if you want to ship an application, you have to write the code, build the container, deploy the application, wait for it to be online-- all these things just to validate that your code works. And it's a lot of things to have to think about, have to go through, especially because as you refine your code, you might have to go through this 15, 20 times per day. Refining that experience so developers don't have to carry all this cognitive load on their head, and instead can focus on their application I think is very important.

And I think that's where the community needs to get involved because there's not going to be a single experience. We all build apps in different ways. We all think of development in different ways. I think that this diversity of thought, of experiences, is very important. And I think that right now, the Kubernetes community is still very Kubernetes-driven, which makes sense. So that's where it's important to start to involve people from different backgrounds who don't care so much about the infrastructure and operations and care more about development.

CRAIG BOX: How is Okteto different from other development tools-- Skaffold, Draft, for example-- that are commonly used?

RAMIRO BERRELLEZA: Our biggest difference with what I was saying before, is that for us, it's all about the application. One of the first things that we want to do with Okteto and we do with the open source project is decouple the deployment of an application from the development.

The first thing that you gain when you are using our open source project Okteto is that you can target any running application of Kubernetes and just swap the application for your dev container. You don't have to worry about how to deploy your application. That can be done automatically. That can be done through Helm, kubectl, whatnot. And then at that point, you have a dev container, and you're just working on that. You contain your code, you can run your process, all without really seeing Kubernetes.

We synchronize our code both ways. That's a way to keep things more efficient for you. And I think that is the biggest difference, is that for us-- Skaffold and Draft are really good at automating all the tasks that you have to do when you're building a Kubernetes app, which is great.

For us, it's more about, can we not even do those tasks? Can we give you, pretty much, a terminal into your Kubernetes application where you can just run your process and restart and hot reload it? You don't have to think of anything else. And that's what the open source project is doing right now. It's resonating with a lot of people. And we've been getting some really good feedback and want to keep doing more things with that.

ADAM GLICK: I've seen that the tool allows you to work with both remote clusters and local clusters. But you're a strong advocate for using a remote cluster instead of running everything on the laptop. Why is that?

RAMIRO BERRELLEZA: Because running a local cluster ties you up to Kubernetes. We want to enable developers to not have to learn Kubernetes to be able to contribute to my applications. It's key. And there are things like Minikube, k3s, Docker for Mac, who make it easier. But you're still running a cluster locally. And then when you think of a company, you may have a hundred engineers. Then you have a hundred clusters running on every machine. It's one more thing you have to think about, that you have to maintain.

For us, that's why I do believe that you have the cloud, it's very easy to run a cluster there, or to get it running. Keeping it running is a whole different beast. Moving there makes it easier. And it's one less thing you have to run locally, and it's one less thing that can go wrong in your local machine.

CRAIG BOX: How do you think about balancing the open source aspects of Okteto with the need to run your business?

RAMIRO BERRELLEZA: That is the multibillion dollar question at this point. It's tricky. There's two things here. For us, open source is very important. We build everything on open source. We want to give back. And I think that the dev experience, design of the experience, needs to happen in the open so more people can show what they think about. They can contribute.

For us, it was very important from the very beginning to do a lot of the things in the open, especially everything that developers use-- the CLI, things they install on their local machine. That part is open source. What we're doing, the way we balance it is we do have to keep the company running. We will have to build products that make sense for companies. What we're doing with our commercial offering is what the company needs to do to enable this vision of remote clusters. Pretty much everything that is dev experience is open source.

Everything that is-- things like policy management, resource garbage collection, self-service access, credential management, those are the kind of things that we're offering as part of our commercial solution. Everything else-- the dev containers, the connectivity to remote clusters-- that's all open source. And we feel like that's a good balance. That's a good way we can give back to the community. We can be part of a bigger conversation while, at the same time, making sure we have a healthy business that allows us to do this work for a long time.

ADAM GLICK: Okteto has a SaaS part of the product where the people go into the dashboard and they have it. But you also have the ability to run things locally through a CLI. If someone is just running the open source CLI components of it, what are they missing out on, besides the obvious enterprise features that you talked about as part of the licensing?

RAMIRO BERRELLEZA: You gain the dev containers. You gain the dev experience. It's a very fast inner loop. What you're missing is what we have on the SaaS, which is the UI base experience. All of our experience we're building where you don't see Kubernetes, that is on our SaaS. We've been able to point Okteto to one repo, and Okteto will just figure out how to deploy that. And whether that's Helm, or kubectl, or a Dockerfile, we just get it running for you in Kubernetes. You are missing that part. But you gain everything else. And that's where, for developers who know about Kubernetes, who are building operators, or Kubernetes infrastructure, the CLI makes a lot more sense. For the rest of us who are building applications, who don't really care that they're running on Kubernetes but they do care that their apps are available and they scale, our SaaS is a much better fit.

CRAIG BOX: Where did the name Okteto come from? Is it because eight is one more than seven?

RAMIRO BERRELLEZA: [LAUGHING] I wish that was the reason! That would make a lot more sense. This was not the first name. We were looking for a name, and we knew that Kubernetes, number 8, and the K is very important. Back when we were all in college, especially in Spain, all the computer programming books were translated to Spanish. And they would translate everything, all the way to words like byte, software, hardware, which some of them are very funny because they had to reuse all these old terms.

So for "byte," they use "octeto" in Spanish, which is really the name for a poem, like an octet, when you have a poem with eight syllables. So it sounds like, oh, octeto. So we swapped the C for a K to make it more Kubernetes-friendly. And we discovered that the GitHub handle was available, and the dot com was available, so we went, yeah, that's what we're using.

And like most names, we didn't love it at the beginning. We were like, OK, yeah we'll do this. I will spend three days trying to name this thing. Now, I'm a big fan of the name, and I like it. And I like that it has that inside story that we can tell at parties and at least to each other.

ADAM GLICK: I love that. Most people try and aim for poetry in their work, and this is quite literally poetry as your work. [CHUCKLING] You mentioned that Okteto wasn't the original name. What was the original name?

RAMIRO BERRELLEZA: The first name of that open source project was CND, because it was "cloud native development." And that's the whole concept that we have with this idea of moving people to remote clusters and to this kind of abstracting Kubernetes away, we call that cloud native development. CND was the first name of a CLI, but then people were very confused. it was hard to type. It was hard to remember. That's when we made the transition to Okteto. Especially as we were thinking of making a company, we thought we needed something more representative of what we were doing.

CRAIG BOX: Changing names is hard. Are there any places where it still shows the old name, or are there any Easter eggs you put in that homage to it?

RAMIRO BERRELLEZA: Yeah. There is this Easter egg that we started using two years ago, and I'm really happy that someone found it six months ago. Most of our demos or blog posts, we made up this fake user, and the name is Cindy Lopez, which-- it's a multi-level joke. Most of it is bad, but it's "Cindy" because of CND, and "Lopez" because "Cindy Lopez" sounds kind of like "Cyndi Lauper." But it's a Latino last name and since all of us are Latinos or Spanish--

ADAM GLICK: Because all you want to do is have fun?

RAMIRO BERRELLEZA: Yeah, it's kind of a wink to Cindi Lauper--

CRAIG BOX: If she was Mexican.

RAMIRO BERRELLEZA: Yes! That's exactly what it is. We even got this Creative Commons of Cyndi Lauper that we put as the profile pic for this made up user. It's a funny wink to all of these things. And I'm really happy that one person got it a few months ago. So I guess that means it's not a super awful joke. It's just bad. But I'll take that.

ADAM GLICK: [CHUCKLES] That's great. Finally, I know you live in San Francisco and love a good Mission burrito, as I saw you tweeting about recently. So I'm going to ask a highly controversial question here, which is, what is the best taquería in the Mission, and which burrito do you recommend there?

RAMIRO BERRELLEZA: Oh, this might get me in trouble here in the city. I may be banned from certain establishments. But when it comes to burritos, I'm a purist. So I'm a huge fan of El Farolito, the one that started the whole thing. Specifically, though, I'm a fan of the El Farolito on 24th Street. Whoever comes to San Francisco, don't go to the one on 16th. Just go to the one on 24th. Walk the extra five, six blocks. It's worth it. I don't know why, but it's just so much better.

ADAM GLICK: And a recommended burrito? If you're going to make the extra steps, you're going to go all the way to 24th, which one should you be getting?

RAMIRO BERRELLEZA: It depends on how adventurous you feel. But I highly recommend getting a beef tongue burrito, which I know for some people it might be squeamy, but it's really, really good. So I highly, highly recommend it. If not, the classic carne asada, you can't go wrong.

CRAIG BOX: Cow in all its forms is fantastic and should be enjoyed at any opportunity.

ADAM GLICK: We'll agree to disagree on that one, Craig.

CRAIG BOX: And with that, thank you very much for joining us today, Ramiro.

RAMIRO BERRELLEZA: Thanks to you, Craig and Adam. This was a lot of fun. I'm really happy that you invited us here.

CRAIG BOX: You can find Ramiro on Twitter @rberrelleza, or on the web at Okteto.com, with a K. You can find Okteto on Twitter @oktetohq.


ADAM GLICK: Thanks for listening. As always, if you've enjoyed the show, please help us spread the word and tell a friend. If you have any feedback for us, you can find us on Twitter @KubernetesPod, or reach us by email at kubernetespodcast@google.com.

CRAIG BOX: Even better, fill in our audience survey, which you can find at kubernetespodcast.com/survey. You will also find transcripts and show notes on our site, as well as links to subscribe. Until next time, take care.

ADAM GLICK: Catch you next week.