Cloud Commute
From Teaching to Tech: The Forefront of the Kubernetes Community - Bart Farrell
When I moved to Spain, I didn't know any Spanish, and I learned Spanish quite quickly. I always felt like I've learned a lot from my students probably than they learned from me, but it was a great way to get a feel for the country that I was living in, the place that I was living in, and also to understand better how information is processed and transmitted. As he said, don't get angry at somebody because the person doesn't understand because it's most probably you not explaining it right. If the person doesn't understand like five times in, either you might want to exchange the teacher or the student, but that is a different thing. The idea with QFM is that our guests are folks that have already taken the time to write an article, so they've organized their thoughts, they've put some structure into it, and what we do is we dive deeper to understand why did you make the choice that you made? What was it that was going on in your head that why you decided to favor this, so that you know better than anybody, they're always going to be trade-offs. You're listening to SimplyBlox Cloud Commute podcast, your weekly 20-minute podcast about cloud technologies, Kubernetes, security, sustainability, and more. Hello everyone, welcome back to this week's episode of SimplyBlox Cloud Commute podcast. This week will be very special. I have a guest that is not like the typical kind of guest we always had. Normally we always have a guest working for a specific company, talking about that product. This week will be very, very special. Thank you very much for being here and maybe just give us a quick introduction on who you are, why you are especially here. First of all, thanks so much for having me. Great to be on the podcast. From looking at the other guests that you've had, I think it's fair to say that this probably will be a little bit of a different conversation given my role. My name is Bart Ferrell. I'm originally from the United States, but I've been living in Spain for the last 14 years. I've been a freelancer for around the last seven years and working in the cloud aid ecosystem for the last four. Very much focused on content creation and event organization. I run a podcast called QBPM, which is part of LearnK8s. Folks might be familiar with that. I've been a CNCF ambassador for the last three years and very active in the community and always possible. Really like interacting and learning from very smart people such as yourself. We got connected in the post-crest world, probably have to talk about that a little bit more in detail, but always excited to learn about new ideas, all the things that we're going on in a very fast paced ecosystem in which we happen to work. All right, cool. Yeah, you already mentioned, we know each other. Actually, not from the post-crest world, which is kind of weird, but from not from the Kubernetes world, but from the post-crest world, that's where we met first time two years ago, last year, something. So for everyone who ever wanted to go to a post-crest conference on Ibiza, there's your chance. It will be in a few weeks, September or something? Yep, September 9th to 11th. Yeah. There you go. You know better. I think you're actually part of the organizing committee, aren't you? Correct. There you go. You do so much stuff in the community. It's absolutely incredible. So yeah, what is the connection to the post-crest world? I think we come back to the Kubernetes world in a second. Very good. So I think the connection to the post-crest world, if I could summarize it, is to be fearless. So in 2020, I had the opportunity to start leading the data on Kubernetes community. And when I started, there were 120 people in Slack, and I went alphabetically and just started introducing myself. Well, it turned out that someone's name started with A, his name is Alvaro Hernandez. He's based in Spain, and we started talking. And so from there, we became friends, and he's very active in the post-crest community, has been for quite some time. And so he was in the data on Kubernetes community to discuss the topic of post-crest operators and Kubernetes. And then I found out that he also organized his own events. As you mentioned, as an Ibiza, they took some years off because of the pandemic. But when the opportunity came up last year to get on the organization team and help out MCing the event and promoting it, I said, "Absolutely." Of course, I'd like to go to Ibiza. I live in Spain for quite some time, but I never traveled there. And we had a great time last year, and we're gearing up for this year's events. You know, with post-crest particular as well, you know, being the battle-tested database that it is, it's been used all over the world, so many different cases. And now also as well as seeing a lot of the work going around when it comes to extensions on post-crest, it's great to see the community that's behind that, all the different ideas, seeing people from all over the world, people that attended last year, all the way from from Indonesia to different countries in Europe, also some people from North America, and this year is looking quite good too. So that's kind of like a connected post-crest. Alright, fair enough. You mentioned Alvaro, and I think I had Alvaro on the podcast in March or something. I'll link the episode in the show notes. It was actually a quite nice episode, specifically because he said he's very big into looking into making operators for Kubernetes, for the post-crest world, and that was the main topic for him, basically talking about StackRes and his own company, Ongress. Okay, so I see your big into communities where the ecosystem is vast. That could be Kubernetes, but it could also be post-crest. It's very anything else we need to know about. That's a good question. I think also, you know, finding your niche in these spaces in the beginning, it took some time, particularly because I come from a non-technical background, so it was quite daunting at first, but I can't stress enough how welcoming everyone's been and how that hasn't been a problem, and really spend a question of what are the other ways that you can contribute. So in different ways, you know, I speak multiple languages, I speak Spanish because I've been living in Spain for quite some time, so that's a way to interact with local communities in Spain. Also, working with, I worked as a teacher in the past, and so, you know, when it comes to giving presentations or things that most people hate doing in terms of public speaking, it's an area where I find myself quite comfortable, and also helping young people get involved in the cognitive ecosystem, having done that in the teaching world as well. There's a very healthy overlap between the two. So yeah, I've definitely touched on a lot of different areas, but I've been able to combine them, whether it's been with music, whether it's been with creating videos, I'm also a big hip-hop fan, so for the data on coronavirus community, I wrote a lot of raps about things ranging from post-crest operators all the way down to, you know, cybersecurity, you name it. I probably wrote a rap about it sometime before the days of chat GPT, I'd like to add, but I've really been lucky that, like I said, that I've been part of such welcoming communities, where everyone is interested to see what other people can provide, and there's no question that's too basic. Being curious is a great way to get conversation started. Yeah, so you said you worked as a teacher before. I love it right on that. What kind of teaching? Good question. So, when I moved to Spain, I didn't know any Spanish, and I learned Spanish quite quickly, but in the meantime, the first job that I could get was teaching English, and so that meant teaching kids all the way down to the age of 8s to adults, you know, in their 60s, and so working at companies, teaching people that work for the government, teaching engineers, teaching teachers, and I felt that I always felt like I've learned a lot from my students, probably than they learned from me, but it was a great way to get a feel for the country that I was living in, the place that I was living in, and also to understand better, you know, how information is processed and transmitted. They say, you know, that if you study the idea of teaching, it's basically knowledge of your subject, and then your ability to transmit it, and so getting angry at students for not understanding something is like, well, perhaps maybe I'm not explaining the best way possible, and so being flexible and adapting to different learning styles, I learned a lot through that process, and also through having to teach teenagers a few times, I've learned how to be patient and focus on objectives. I think also it doesn't matter whether it's, you know, whether it's English, or whether it's Postgres, the best environments, you know, in terms of teaching is when your students teach each other, when you create an environment where they're able to share knowledge with each other, and you're just more of a facilitator, so yeah, so I did that for about four years, and at the same time then I started helping companies that would be giving presentations internationally, with repairing that, so focusing more on public speaking, and if you continue to do that also with the entrepreneurship entrepreneurial community in Bilbao, in the north of Spain where I live, and that's also carried over to the Canadian world, where I work as a consultant, so every CEO, every CTO needs to go out there and be able to explain their ideas, and needs to be able to do so in a short period of time, that's practical, focusing on the value of what it is that they're offering, particularly considering that in the world in which we work, we're talking about things that are generally not tangible, so how do you make something visual, how do you make it come to life, and particularly with all the competition that's out there, how can you really double down on the pain points that the end users might be facing, so that they see that it's actually going to be providing something that the current alternatives simply don't. That is interesting, because as he said, we're working mostly in content creation these days for the cloud native community, or the Kubernetes community, most specifically, and as he said, I think teaching people will give you a good insight into how to get a specific topic across, how do I have to explain stuff, and as he said, don't get angry at somebody, because the person doesn't understand, because it's most probably you not explaining it right, I mean, if the person doesn't understand like five times in, either you might want to exchange the teacher or the student, but that is a different thing. I mean, I had it in the past, but literally a second person came in, just phrased it slightly different, and it literally made click for the person, so sometimes it is this like, okay, I'm just not the right person. So from the CNCF community itself, you're traveling a lot, you're doing your own podcast, you're doing these short interviews, specifically for the Kubernetes, that was like a massive series of shorts. How many did you record for that? About 30, and actually still playing on doing more throughout the rest of the year, but the core focus was, you know, since June was the celebration month, that's why I stopped the conversation there, we got a lot of interesting feedback from those interviews, and a lot of it, we asked two simple questions, so is what's your least favorite feature on Kubernetes, and so we got lots of different things. Some people still actually writing stateful workloads with something that was mentioned, so you know, the database problem, you know, working with data is still tricky for a lot of people, but by far the one that seemed to stand out the most was networking, and you know, managing network policies on Kubernetes seems to be very frustrating for a lot of people, and the other question that we asked was, you know, the future, what do they imagine will happen in the next 10 years of Kubernetes, so a fair amount of people are responding with the idea that Kubernetes becomes a platform for creating other platforms, Kelsey Hightower is referred to, that it will, you know, simply blend into the background, become a lot more boring, if we look at, you know, the history of Linux, what can it teach us about where Kubernetes will be headed as being a well-adopted accepted, you know, you know, technology that across the world is just, you know, de facto standard, which, you know, de facto standard we can say for container orchestration is already there, but continuing even further its adoption, becoming less of a Wild West and more something that just very much standardized. So those were good, then the main episodes that we do for QBFM we find engineers that have written blogs and we bring those blogs to life through a conversation, and so going back to the idea of teaching is that everybody learns in different ways. So for example, in your case, what works best for you to learn? Do you watch YouTube videos? Do you read blogs or read books? What's your what's your go-to resource? I started programming when it was like 10 or 11 years old, so at that point in time it was all books and trial and error basically, right? I mean, there wasn't a lot of stuff around it. I'm probably older than most people expect, so that was like 30 years ago, just for the ones that really wonder. YouTube was not a real thing at that time, not even, not even the beginnings. These days, I still think the like trial and error works best in terms of like that you best understand how stuff works and why it probably doesn't work, but I think to be fair, most of the banks is probably like YouTube videos and stuff like that. I have a good chunk of books. It's very, very hard to read books these days. Still not a TikTok guy, but YouTube videos or in general, like video courses, I think are the current like way to go. Okay, that's great. In this case, you know, with, and like you said, whether it's attention spans, you know, whether it's the influence of TikTok, whether it's the influence of, you know, YouTube shorts, et cetera, et cetera. A lot of times I've got to go. I want a very specific piece of knowledge to give it to me in the shortest period of time possible. Like I said, everybody learns differently, so but the idea with QFM is that our guests are folks that have already taken the time to write an article, so they've organized their thoughts, they've put some structure into it. And what we do is we dive deeper to understand why did you make the choice that you made? What was it that was going on in your head that why you decided to favor this so that you know better than anybody, there are always going to be trade-offs. If I use this database or that database, what am I getting? What am I gaining? What am I losing? We're talking about, you know, tools in the CNCF ecosystem. If I use flux, as opposed to ARGO, we're going to be the trade-offs there. Is it a question of maturity? Is it a question of my use case? Is it a question of the community backing it? In some cases with CNCF projects, one could be sandbox, one could be graduated. Then we can even talk about, you know, open source versus, you know, not like all the different things that are going to go into that. So the conversation is designed around, you know, helping our audience understand the decision-making process that the guests went through. A lot of times as well, too, you know, having, you know, sort of horror story situations of when something went wrong, this is what I did. So let's face it, things are going to go wrong. There are going to be fires in production. There are going to be outages. Things are going to go down. You're going to have to deal with that. So how do you... Part of it is the, we can say, the emotional intelligence side of it, of keeping calm skills that are not often, you know, taught, let's say, in universities or, you know, in every context. So combining those skills with then the more technical ones, that's kind of what we seek to do. And once again, with the idea of educating people of saying, look, you're possibly going to be in this situation. If you are, this is how I handle it. You may decide to handle it differently. And so doing that with the technical focus, and it's always vendor-neutral content, it's just engineers helping out other engineers. That's the idea. I mean, that is a good point, right? I mean, there's a reason why these, like, lessons learned, best practices, whatever you want to call those talks are so big at conferences, because it is like the best content you can probably generate in the sense of like, okay, why did I do something? Or what was, as you said, what was the decision-making? What was the thought process behind that? And I think the thought process in itself, like, evaluating certain things and making a decision why stuff works or it doesn't work or would not work in your case, I think that has been an interesting thing. And the other thing you mentioned is that stuff will always break. And that is a good point, right? It's important to make people aware of that. In the past, I did talks on fault tolerance, and that it is always a trade-off between how much money do I want to spend and how much money do I lose if my system is down? And, right, there's no way to build, like, 100% fault-tolerant failover system. You will always have issues. The question is, how much money do I lose per second if I'm down? And what is the trade-off I have to put in money to make it not happen? But because you said, one of the questions of your Qubevam podcast is, like, what's your least favorite feature in Kubernetes? Let's flip the ties, and what's yours? Oh, I would say, so I would say, well, that's a great question. And what I would say, and I do agree, this was mentioned by a couple of people, is the vast complexity just to get started with Kubernetes. I think that it's still something, and as a CNC ambassador, it's something that I look forward to working on with other ambassadors, so that not just for non-technical people, but so that the process is less overwhelming. Because you see, like, plenty of memes where it's, all right, you know, you imagine you're on the sort of stairway to heaven towards, you know, the guiding light, the shining light of Kubernetes, or that, and, you know, that stairway collapses if you fall into a hell of all these different kinds of technologies that you have to learn. So I understand this, you have to understand that. And I think it's a process that a lot of people go through, which is why that learning curve is still quite steep. And inside the ecosystem, I think we get the feeling that you're going to walk down the street and you ask anybody what Kubernetes is, they're going to have a perfect answer for you. And it's simply not the case. You know, it's a, and maybe it's because of living in Spain where the adoption, you know, is a couple of years behind, if we look at the US and we look at other countries in Europe. But I think, you know, you talked to a lot of folks out there that are still running on very much, you know, monolithic legacy systems. And when you talk about Kubernetes, they know that they want to make the jump. They know that they want to do it. But what's the first step? How can that be easier? So I would, I would say that as someone that really struggled, particularly going in directly into a sort of even more niche area through the data on Kubernetes community, because running databases on Kubernetes for a lot of people still sounds like crazy science fiction. That was particularly challenging for me, and that was quite painful. But through the patience and kindness and guidance of folks like Alvaro and lots of other people that I've met along the way that made my journey less painful. It was, it's been easier. I guess you can kind of say it's the sort of, you know, it takes a village to raise a child. And I was that child and very much still in that child. So I would say, yeah, the complexity of getting started. I'm not discounting the massive amount of work to spend on the cloud native glossary. There's been a lot of initiatives that have gone that people are still putting a lot of work into. But I feel like a fair amount of it still assumes a certain, a baseline that for me is quite high. But once again, as a non-technical practitioner, it's only natural that seems that way. So yeah, I would definitely, I would say that. I think that makes sense. I mean, I'm coming from a very technical background. As I said, I started programming like super, super early, and I still found the, the get first steps into the Kubernetes world kind of complicated. So I get what you're saying. And the other thing that I personally think is, and that goes for any kind of abstraction layer, which Kubernetes from my perspective is, we call it an orchestration, but in general, it's, it's an operating system abstraction. Everything is awesome until it breaks. If it breaks and you have no idea why, try to find it. Good luck. That is, that is my personal biggest issue right now. We used it in my own startup and it, it was awesome when it worked. But the few times where you had an issue and it all didn't make sense. And that, that is where also the ecosystem comes in, right? Then you have, like, as you said, Flux, Argos CD, you, you might have some, some service mesh or whatever. And those things all just make it more complicated, unfortunately. They just make the, the learning curve even steeper than it has to be. But apart from that, yes, Kubernetes is, is, is, is really good. I think one of the main benefits is that, and, and we have the same thing with Docker in the beginning, if you, if you run Kubernetes, like even mini cube or k3s or something on your, on your machine as a developer, it still brings you closer to production, right? This, like, as a developer, like, I made the commit. Now it's, now it's the operations, the operating teams problem. That is not a thing anymore. And it makes it much harder to, to find out why certain things don't work or why they break in production. And Kubernetes makes that much easier from my perspective. We already crossed with 20 minutes, Mark, I can't believe. Let me see, but there's most probably one, well, there was certainly one other thing that I, that I wanted to bring up. Oh, right. The, the databases on Kubernetes and in general, stateful workloads on Kubernetes. It's, it's a very good point. I mean, with simply luck, we're in the same, same world. We believe databases belong on Kubernetes, at least, but ones that are mission critical. But we see the same problems. We see a lot of people facing us like, no, stateful workloads are not, or Kubernetes is not designed for them. Like 10 years ago, or like, like five years ago, I would have a great, maybe not 10 years ago. But five years ago ago, I would have still agreed with you. But a lot of stuff changed. The CSI driver interface is, is absolutely incredible for people that probably don't know if the CSI is the container storage interface. That, that thing is, is inherently designed for stateful workloads. Otherwise it wouldn't make any sense, right? So I think that is, that is a very good point. And the data on Kubernetes community is, is a big step forward, making sure that people understand that. We haven't joined yet, but it's, it's on, on our list to do still. Oh, good. I'm no, I'm no longer leading it. But there are some really talented folks that are, that are interacting there. And, you know, there's always presence that could have gone. I think also, it goes back to the, the pain point, right, is that when in doubt, if you're struggling with something, find a community, you know, make community part of the solution. And so whether it's data based on Kubernetes, or it happens to be any other technology, there are other people out there that have probably gone through something similar, leverage that. Yeah, for, for, for us, I, I started joining the CSI or the special integration group storage calls a while ago, and then trying to be a little bit more active, because it, it's, it goes both ways, right? You're, you're leveraging the community, but that means you all should, should give back to the community as well. All right. And because of the 20 minute mark, last question, the question I always ask, well, that I always ask people, what do you see in the future? What do you think is like the next big thing in your case, Kubernetes, VCNCF, databases, data on Kubernetes, AI, whatever you can think of? That's great. I mean, from the last KubeCon, and particularly having, you know, when I was writing the data on Kubernetes community, we were talking about AI and ML workloads becoming more apart on there. There's a great book by Patrick McFadden and Jeffrey Carpenter about managing stateful data on Kubernetes, sort of Riley book that they wrote, and they do talk about those workloads at the time, not only at the time that the book was written, but it was just seeing that, oh, that's kind of further up the stack. Now, perhaps it's because AI hype waves, well, I would say there's a significant portion of the, that is related to AI hype waves, but that the cloud native world feels like, oh, we can't miss the boat. You know, we got to catch up. We want to make sure that we're on top of this. We saw a lot of this in the last KubeCon. And so what I'm, what I would kind of, you know, ask you, and also I want to hear more from platform engineers, SREs that are out there, how much are they using AI in their day-to-day work? And, you know, in the sense of what value is really being provided earlier this year attended event in the UK, where they spoke a lot about that, how many, you know, how many startups are really, we can say, running AI, you know, doing AI legitimately on their own, rather than just running everything back to open AI, I want to see where this can actually move forward in practical terms. Or is there, is it still, is there a lot more time that's needed for these two worlds to come together? I think that's, it's certainly a question that, that came up a lot, and I'm not seeing not many concrete answers yet. So it's a question that I'm going to keep asking. So I don't know if you have any thoughts on that, or it's a conversation that you can continue the future. I think it is a conversation that needs to continue. I'm pretty much on par with your opinion. I personally don't believe that every single tool we use needs an AI agent or any AI feature in general. And I think that is part of the hype train you just mentioned. I guess it will, like, slow down at some point, and only the use cases that are valid will stay. But apart from that, I think it's interesting. And because you earlier asked how I learn, I literally just published yesterday a blog post on image recognition within your networks, and I literally knew nothing about it, except for it somehow works. And I literally used YouTube videos and a blog post trying to figure out how that stuff worked. It was quite an interesting thing from the idea of like chat GPT or cloud AI or whatever. I like to use those to not start on a blank page, not generating the necessary content, because I think it's not a good idea to ask an AI tool or an LLM tool for its knowledge, if you can't pass it in at least, if you don't pass in content. But I think it's great to generate an outline. What are the interesting topics you could investigate and write about? I think that makes a lot of sense. Anyway, we just crossed about 26 minutes. So thank you very much for being here. It was a pleasure having you. We see at a few conferences in the very near future. I guess you'll be in Ibiza. I guess you'll be in Salt Lake City for KubeCon. There you go. And we'll probably see a few more conferences in between as well. For the audience, thank you very much for being here. I hoped you like this slightly different format and the usual stuff. If you like it, let me know. And for the ones on YouTube, subscribe, blah, whatever. Thank you very much for being here, Bart. My pleasure. Thanks for having me. The Cloud Commute Podcast is sponsored by SimpliBlock. Your own elastic block storage engine for the cloud. Get higher IOPS and low predictable latency while bringing down your total cost of ownership. WWW SimpliBlock I/O.