All Things Docker

Justin: [00:00:00] Well, let’s go as soon as if isn’t quite everything on the planet. Sometimes it feels like it.

Bridget: It’s time for Arrested DevOps, the podcast that helps you achieve understanding, develop good practices, and operate your team and organization for maximum DevOps awesomeness. I’m Bridget Kromhout. We have a great show for you today about all things Docker. But first, a word from our sponsors. This episode is sponsored by CircleCI. Designed for modern software teams, CircleCI’s continuous integration and delivery platform helps developers push code with confidence. Trusted by thousands of companies, from 4-person startups to Fortune 500 businesses, CircleCI helps teams take their software from idea to delivery quickly, safely, and at scale. Visit arresteddevops.com/circleci to learn why high-performing DevOps teams use CircleCI to automate and accelerate their CI/CD pipelines.

[00:01:04] This episode is brought to you by cloud-native consultancy Container Solutions. We bring culture, strategy, and technology together to help you get your cloud-native transformation right. To find out more, visit arresteddevops.com/containersolutions. This episode is brought to you by MacStadium, leading provider of cloud solutions built on Apple Mac hardware. As more teams are working from home, having your Mac build infrastructure in the cloud can make it easy for your app devs to work more efficiently. No need to have someone in the office keeping an eye on the Macs. Let MacStadium do it for you. And if you need a fast, scalable, modern way to run Mac virtual machines, MacStadium’s virtualization platform, Orca, is purpose-built for running iOS and macOS CI. Orca takes a standard macOS VM, puts it inside a Docker container, and then uses Kubernetes to orchestrate everything, all on Mac hardware. Orca is easy to integrate into your current workflow with plugins for all the popular CI tools like Jenkins, GitHub, GitLab, and Buildkite. If you’re building apps for the Apple ecosystem, learn more about MacStadium at macstadium.com/arresteddevops. From that link, you can also get access to a free 2-hour sandbox to give Orca a try.

[00:02:28] Joining us today are 2 amazing guests. First, talking to us from the UK, uh, we have Justin Cormack, CTO of Docker. Justin, what do you want to tell us about?

Justin: Hello, it’s lovely to see you.

Bridget: So our other guest, we’re lucky enough today to have Donnie Burkholz, VP Products at Docker. Say hi, Donnie. Hey, everybody. So let’s get into it. First, you folks had your Community Day yesterday. It was actually kind of funny. I wanted to schedule this podcast and then, oh, hold on, they have a Community Day at that very moment. So you wanna kind of tell me what did we hear about at the Community Day? What’s going on with Docker in 2021?

Justin: It was actually also, it was our second kind of Community Day. It’s been weird not having in-person events and things. So we started doing these online things. We had 2,500 people turn up, which was great. And we, We talked about what Docker’s about and what we’re planning to ship and what kind of things we’re working on. I think we— and we showed some little demos and things of things we’re working on, experiments that we’re working on, and ideas we’re having to get feedback on. Donnie, do you want to talk about some of those?

Donnie: [00:03:57] Yeah, absolutely. And I joined Docker about 5 months ago. And I would say that one of the upsides of COVID from my view has been the increased virtual-first and remote-first nature of the world. It gives me an opportunity to be in Minnesota and still do really cool stuff with cool companies. And, you know, so Docker being a company I regard as super cool, you know, some of the things that I’ve been bringing into them was trying to help us focus on customer problems. I mean, I think a challenge that a lot of tech companies have is like building very solution-out of like, here’s the thing I’d like to build, and I’m going to build it, and then I’m going to ship it. And then, you know, inevitably, 3/4 of the time, it flops because you don’t really know what people’s problems are and how you’re trying to help solve them. And so a big focus for me has been, how do we make sure we’re listening to people who are using Docker, people who are paying for Docker, synthesizing the biggest pain points they have in their development workflows, and then focusing on solving those in ways that really feel magical and maintain the love that a lot of people do have for Docker. Some of the things that we hear a lot about are things like developer productivity, developer velocity, finding stuff that I can use that I can trust instead of just random repositories somewhere with 3 stars on GitHub that seem to be the only thing that’ll solve your problem or whatever somebody copied and pasted on Stack Overflow the other day. Those are— it might get the job done, but in a way that everybody agrees you feel a little bit nervous about, like, should I be shipping that into production? Not quite sure about that.

Bridget: [00:05:43] It’s that question of like, do I take a dependency on this? It hasn’t been updated since 2019 and maybe the panini or the Panera caused people to not update it, but maybe this project is never going to be changed again. Um, but when people are trying to figure out like what’s going on, I think for Docker specifically, and I’m going to start by directing this question, um, to Justin, I feel like a lot of companies don’t necessarily put a public roadmap out there. You folks have a lot of transparency and you do. Can you talk a little bit about what people should expect if they want to either get involved with some of the open source or they’re taking a dependency on some of this and they’re looking at it and thinking almost there, what does that mean? Developer Preview Experimental, is that dangerous? Like, what is that? Can you kind of give us a little guide to what’s going on with the roadmap there?

Justin: Yeah, so we try to be very clear about, you know, what, you know, whether things are ready or not. And we’ve, I know we tried to change a lot last year about how we do this as well, because previously when we were shipping enterprise products, we had this, thing where we labeled everything very much as experimental, do not touch. We would only turn things— we would only turn— you had to figure out the magic runes to how to turn something on so you could even try the experimental thing. And that really made, you know, there was a very anti-trying things out thing. It was like there was this culture of whenever we’ve shipped something, it’s until we know it’s right, it’s marked as experimental. And when we know it’s right, then we’ll take the experimental flag off and then everyone can actually, can use it. But of course, that’s not really how things work because most people don’t try the things that have the experimental flag ’cause either they can’t figure out how to do it or it’s like, oh, this is labeled as dangerous. So we’ve really been trying to cull all the kind of experimental, turn it off as experimental things and go for much more like maybe a little notice or something saying that something’s new all these statuses here of like, we’re trying it out, would you like to experiment? And we’ve got, you know, different groups of people. We, you know, we have developers who would like to try things out and ask to be fed the new stuff. And we’ve got people who are, you know, maybe slightly later adopters who try them when their friend recommends them. So the status we’ve got on the roadmap is, first of all, for something to get on the roadmap at all, we put some things on the roadmap because we’re thinking about them and we’d like feedback, but anyone else obviously can put things on the roadmap as well. So if you’ve got an idea or something is bugging you or whatever, like don’t hesitate to open an issue. And so that’s the kind of, you know, that’s the state, you know, and we look at them, the product managers look at them every day. We have a weekly meeting where we go through and look at what’s new, what’s been voted up as well. We look at the, We look at the thumbs up votes to kind of rank, like, has something got a lot of interest this week? Is there a lot of people saw this and thought, oh yes, I would like that too. Or is it just, is it something that’s quite, you know, that’s just one person’s idea so far? And then things sit in, like things go to investigating because we’re like trying to work out, is this feasible? Is this possible? How could we do this? What could we do? So that’s that status. Then writing the code is when We’ve pretty much decided, at least on first principle, what it’s going to look like. Almost there means we’re probably asking people to try it, but, you know, behind maybe the, you know, the early users, the people who raised it in the issue and things like that. And then developer preview experimental means there’s some public way of trying it. So there’s more feedback and then shipped and joy means it’s actually gone to some sort of GA. But obviously we still potentially iterate on that or raise an issue if you need to change it. But you can also, at that point, it’s where you can raise direct issues on the actual product area or or things like that because it’s in product feedback or whatever, and we can see if people are using it and so on. That’s the lifecycle there. We’ve got a lot more ways that we’re trying to get things out early to people and without these experimental flags and blocks on people using them.

Bridget: [00:10:36] That’s great. That’s fantastic. I’m sitting here looking at this thinking, great. We’ve now heard from the CTO about all the exciting technology. Donny, you have a focus on product as a VP product. Can you kind of look at this and tell us, well, what’s a product? Product, project, what’s happening there?

Donnie: Yeah, I think that’s, you know, something that Docker as a company, we continue to evolve our view on that as well. You know, I think there’s a long history of giving lots and lots of things away for free and being an amazing community and amazing open source technologies and amazing free SaaS services and not so much of how do we turn this into something that creates a really sustainable business model so that Docker will be around in the future and will continue to thrive in the future as it combines what it’s done and where we’re going. When I think about this in the context of our roadmap, for example, one of the things that I’m looking for when we put something on investigating is, What’s the level of interest in us doing it at all? Is it valuable for people? Is there some kind of a way that we can turn that into something where people want to subscribe for Docker so that they can get access to, you know, an additional level of capability associated with some of those problems they’re trying to solve? So it’s not only about the technical feasibility part of it, it’s also about, you know, is this a valuable pain point or problem for people? Are there you know, as Justin said, are there hundreds of upvotes, as we saw, for an example, with, uh, the, uh, the M1, uh, release, right? In the span of days after Apple made their announcement, it turned into the most upvoted item on our roadmap of all time.

Bridget: [00:12:18] I’m really interested in that because how did that happen so fast? It’s not like everyone had that hardware in their hot little hands immediately. Like, how do you, how do you see that as being like aspirational versus actually actionable? What do you think?

Donnie: Yeah, it’s, to me, it’s been a, it’s an interesting path of like abstract to concrete because everybody knew a long time ago that Apple was going to be doing something ARM-like on some of their hardware. And it was a question of like what and when and will it be like the developer versions of hardware, just kind of the consumer-grade editions that don’t have good CPUs and not enough memory and whatnot. And so then they announced, then they said, hey, we’re gonna have these Pros and they’re you can like buy them now. And so instantly people turned around and said, oh, but my toolkit is not available. Like none of the tools that I need, right? Docker being one of them, but none of them work, or at least enough of them don’t work that I can’t use this new toy that I wanted to purchase. I’m stuck.

Bridget: [00:13:19] So everyone is suddenly saying, hey, I need Docker Desktop and I need it 5 minutes ago. Like, does that— when Apple kind of drops a an apple-shaped zero-day on you, as it were? How do you, as a product organization, how do you react to that?

Donnie: It’s tough. And I think, especially considering Docker is not a huge company, and so we have to be really careful about where we put big engineering investments in terms of making sure we get the customer signal first, instead of just making wild guesses and hoping they work out. So this is one, you know, where, yeah, we could have done it earlier, but we didn’t have a great set of signal on like what is the timing and what are the different alternatives of things where we could be putting our time right now compared to where we might want to put time in the future. Because it’s like a lot of technology, you know, you know it’s going to come, you know it’s going to happen, you just can’t tell when.

Bridget: Yeah, yeah. And so there’s obviously when we try to either look at our specific product roadmaps or our hardware vendors’ roadmaps, or in the case of Justin, you could talk a little bit about maybe even the CNCF landscape because you’re also in leadership in the, I believe, the TOC, the Technical Operations Committee? Oversight. Oversight, that’s right. Oversight, yeah. So many acronyms. Technical Oversight Committee. Can you talk a little bit about that? How you make technical decision-making in a space where the scope is apparently everything on the planet?

Justin: [00:14:54] Well, the scope of CNCF isn’t quite everything on the planet. Sometimes it feels like it, like it’s the— but it’s everything. I mean, it is a very ambitious— cloud native is everything that is modern software and modern software delivery. So it is a huge space in that sense. I mean, the advantage of it to some extent though, you can see things coming, you know, and there is a community that, I mean, there’s a huge community around CNCF and cloud native that really, you know, is very open and we talk to each other a lot and we know what’s coming most of the time. And I think that sometimes from the outside it looks like, you know, sometimes like there might be a consortium of people come together and suddenly announce something, but actually almost, you know, a lot of people know about these things beforehand and they talk about them. And there’s actually much less of that whole kind of big announcement culture than there used to be. And there’s much more of the, you know, let’s collaborate on something we’re working at. We’ve got this idea, let’s work on it together. And that kind of thing. And so there’s a lot more early signal about the kinds of things that people think are interesting, the kind of direction they’re going. And yeah, there’s, you know, there’s bits and there’s a very, there’s a big, you know, consolidated ecosystem around everything you need to make Kubernetes work in CNCF. And, you know, over time we’re discovering it takes more and more things to make it work. And then you’re looking at anything, well, actually, yes, that really is, it’s the container ecosystem of everything you need from, you know, development to production. And so far a lot of CNCF stuff is around the production side. End of things, but there’s a lot of stuff floating around in more developer space as well. Like the— I’ve temporarily forgotten the name of the Spotify project that joined CNCF recently, but which is around much more around user experience for that. And there’s a lot more people thinking about developer ends of things and not just infra. And there’s people thinking about how we build applications, which is obviously an area that Docker’s really been involved in for all its life, really, that kind of area. And so all these things are starting to come together and there’s cooperation across all these areas. You know, sort of in the sense of TSC is really just looking at things and assessing how mature they are, talking to people who are using them and finding out if they’re working for them, which is a really valuable thing to do. I really, I really like, you know, assessing projects because it means I just go and I can go and talk to people who are using them and find out how they’re using them and if they’re working for them and what kind of problems they have, which is a really nice thing to have an excuse to do. I mean, I would probably do it any, you know, I’d probably do it anyway, but it’s like I’ve got a reason to talk to you so I can go along and ask you how you’re getting on with you know, with buildpacks today, you know, is it working for you? It’s like, how are users adopting them? And those kinds of things, which are really, you know, interesting things to get different views from all the different parts of the ecosystem.

Bridget: [00:18:18] I’m curious, and I want to hear what Donnie thinks about this, how the insights that Donnie, of course, has a background as an analyst and an executive in other, you know, end-user companies, you have this background of, you know, you’ve been a technologist for a long time and have insights into a lot of projects that aren’t just the ones in your remit. I’m wondering if, Donnie, you want to address how seeing what’s happening in the ecosystem is informing which things you open source, which things you make product decisions about.

Donnie: Yeah, I think, you know, the first part of that where Justin and I really overlap well is when we can have conversations about customer problems. Because we can bring all of those different perspectives to bear on that, whether that’s from the technology, from the community, what’s going on at different vendors or competitors kind of across the technology landscape. It’s really what are the problems that we’re trying to solve, which ones are unique to us, which ones are broader, and then, you know, from the view of like what do we open source, where do we use open standards, and where do we decide to keep things as our our own products, it’s really about what do we want to accomplish, what is the maturity of that particular level of the stack, where are we trying to differentiate compared to, you know, where do we feel like it’s increasingly becoming, you know, more of a utility or commodity, right? Because like open standards are a great place to help align a broad set of vendors around things that are starting to become a commodity to make sure that things work consistently for people who are trying to use the software wherever they go.

Bridget: [00:19:55] And I think I see a lot of end users, whether it’s through acquisition or through a variety of departments, they end up using a lot of different, not always interoperable technology. And any place that they can get interoperability, any place they can just ship that container, they are happier, that is for sure. Okay, so I want to pivot and talk a little bit about the fact that Obviously, we have Justin coming to us from, you know, Cambridge. I believe you’re located in Cambridge, right?

Justin: That’s right. Yes.

Bridget: Yes. And Donnie here in Minnesota, and people are checking this out from around the globe. I’m wondering if what you’ve seen, both of you have seen different this year in terms of, you know, hiring, onboarding, educating, and reaching your communities. Because obviously you were both completely involved. You were extremely online already, completely working remote, or at least collaborating across the globe already. What stuff has surprised you, or what have you seen that’s different this year? I’ll start with Justin.

Justin: [00:21:02] I mean, Docker made a decision pretty early on that we would become fully distributed and not have offices anymore. We were It’s interesting because I’ve been at Docker for 5 and a half years, and over that period, we’ve gone from being a very centralized in San Francisco company to a very distributed company. It took, you know, that journey took quite a long time, but we were already, you know, well on that way. You know, we were, I think, half European, Um, by the end of 2019 in terms of where people were, and, you know, half US, half Europe. So we kind of become— we’ve become pretty distributed and we’ve become much better at working like that. But we, you know, it— and so it seemed quite, you know, quite natural to do the let’s not, um, let’s not have offices anymore thing. It was, um, it was less of a jolt and like, you know, it was like kind of— it seemed I mean, the lease on our Cambridge office expired maybe 6 months ago, and it’s like, it seemed kind of weird to renew it at that period and things like that. So I think that we were perhaps further on that journey than some people. But I mean, obviously there’s been people who started off fully remote and it’s kind of becoming more normal, but we were well on that journey. So we’d been doing things like hiring people pretty much regardless of where they were before, like Donnie. But it’s become, you know, it’s just become, you know, it’s not, we know we don’t really ask people anymore, particularly where they are. And, you know, it’s become much less relevant and the onboarding experience I think that’s for, you know, people starting has changed a little because again, we used to send people, we used to send everyone to San Francisco for a week, but actually that was becoming less and less relevant as people actually onboarded, you know, in France or the UK and didn’t actually, there wasn’t really much point in sending them. It was a kind of a relic of the centralized culture that was one of the things that took longest to go away, I guess. So that kind of change really just made it— it made the change happen faster last year than it would have done otherwise. We probably would have been there in a couple of years anyway, I would think, or maybe even last year.

Donnie: [00:23:50] Yeah, I mean, it’s been the same as it has been for a lot of companies. I think Justin’s point at the end there about accelerating the pace of change is extremely true in terms of the way people are working internally and the way they’re interacting with others outside of companies, you know, at different communities. Like, we’ve already seen, you know, an interesting blend of, you know, in-person events for local communities and then, you know, some combination of national conferences, regional conferences, global conferences. But a lot of people can’t access those for time or budget. And so, you know, there have been other, like, online communities on Slack or, like, different even webcasts and that kind of thing, right? The whole intent of those is to reach people where they are. Instead of making them come to you. And so I think this has definitely accelerated a lot of that. You know, it’s driven a lot of people using online platforms that are much better suited to online events, instead of trying to force-feed synchronous in-person models into online. And like, it’s the same as trying to force-feed, you know, models that are optimized for one office into a global distributed company, and kind of watching them fail when you’re like, oh, well, there’s only 1 hour of overlap every day between Germany and San Francisco. So how do you make that work if you rely on only synchronous ways of working? And the answer is you can’t if there’s any collaboration involved, right? You either optimize for autonomous or optimize for asynchronous or both.

Bridget: [00:25:17] Yeah, well, I— and I looked and I saw that Docker has a whole bunch of, you know, career openings available if people are interested that definitely are very remote first.

Justin: Yeah, absolutely. Yeah. And, um, it’s, um, yeah, we, we really are just, you know, we really don’t mind where people are. We, um, I think, you know, people are actually moving around, you know, people are moving around in the US and we’re not sure we’ll even know where they are anymore at some point.

Bridget: I mean, tax purposes, that’s pretty much the only thing, right?

Justin: Yeah, I think the US has hiring restrictions in different states because I remember filling in paperwork for one state at one point.

Donnie: But yeah, it’s taxes, it’s hiring, and, you know, at the level of countries, it’s being able to kind of legally operate in different places. And that’s the main limiter besides time zones. Like right now, as we think about where might we be interested in being able to additionally hire, it’s really about how do we enable the level of time zone overlap that we want to have right now. And whether that’s specific geographies that already overlap well, or whether that’s looking for people who are interested in working, you know, different shifts who live in places where the time zones don’t overlap quite as well.

Bridget: [00:26:35] Okay, so I want to wrap with a discussion of the upcoming DockerCon. So I know that the DockerCon Online CFP closes in just a few days here. I saw that there was a blog post about how to write a great talk proposal for DockerCon Live 2021, and I’ll put a link to that in the notes. I just want to hear from both of you, starting with Justin, what are you excited about coming up, whether it is stuff that you already have on your roadmap or stuff that you know customers are doing that they’re going to tell you about? What are you looking forward to feature and to see at DockerCon and in the year ahead?

Justin: DockerCon last year was one of the first online conferences that happened because we planned it to be online anyway before the pandemic. So we were kind of more prepared than other people and it had an amazing attendance. I think there were 80,000 people there. It was really, it was gigantic. And I mean, it really is one of those things that like you, it makes you realize how much conferences are more, so much more accessible to people when they don’t have to pay to travel to somewhere in the US and they can just attend from home. And that was really amazing. And really, I’d really like to see people’s, you know, stories who wouldn’t otherwise be heard because they wouldn’t attend. And I think, you know, we always had a sponsorship program and things for DockerCon, but that was only a small handful of people. And I think that just, you know, developer stories from from wherever you are and whatever you’re doing are really important and we’d love to hear them. There aren’t that many big conferences that really focus on like general developer topics and DockerCon really is like that. People always ask me like, is DockerCon like KubeCon or something? Like if you look at the people who attend, DockerCon is maybe 80% people who identify strongly as being developers, whereas KubeCon is, quite mixed, but like the strongest contingent is people who work in infrastructure. And so it has a very different like feel from the audience point of view and the kinds of things we want to— people, you know, we’re interested in is people who are developers and are, you know, trying to get stuff done. And, you know, maybe, you know, are you using Docker and it’s helped them to do things? And those are the kind of stories we’re looking at. And really, I mean, I think the, um, like, if you’re not sure if the thing you want to submit is the right kind of thing, then if it’s for develop— if it’s interesting for developers, then submit it. Really, it’s, um, like, don’t be too— don’t look at the guidelines too narrowly. But I mean, we tried to, we tried to ask people to talk about experience of working together in teams and, and those things that we’re, you know, trying to— we’re trying to help from the product point of view is work people to collaborate better, um, and in teams and, and work together, because that’s always been, you know, an important part of the thing, of the learning experience and things, is like, it’s helping your teammates on board, help it, you know, helping them use the tooling, helping that, making, you know, images that you can share between your teams so you can develop more easily and those types of things. And so that’s why we put that team collaboration in the CFP, but we’re really interested in everyone’s experience, you know, as a developer of modern applications and what they’ve learned.

Bridget: [00:30:12] All right, what do you think, Donnie? What’s on your horizon DockerCon-wise and coming up this year?

Donnie: Yeah, I mean, you know, I think the thing that really makes conferences valuable for people who are going there is being able to swap stories with each other, to learn from each other about How do you, how do you make yourselves better? How do you make the way you’re using your software better? And how do you, how do you use that to get your job done? Because nobody wakes up saying like, oh, I want to use this tool today. They wake up saying, here’s what I want to get accomplished today, and here’s the things that I’ll need to help me get there. And I really want those things to be, you know, amazing and magical to use and stay out of my way and let me get stuff done, rather than being constant kind of frustrations and I just saw a stat the other day, and I honestly don’t know how real it is, but it concerned me, which is somebody saying that like 50% of developer time on cloud-native applications might be spent configuring stuff rather than writing net new code. And I’m like, well, I wouldn’t disagree, but I do want to learn more to see how real that may be.

Bridget: [00:31:17] I wonder how much of that has to do with usability and how much of it has to do with security compliance, uh, meeting the requirements you need to meet. Security shouldn’t be about configuring things unless you’re filling out a form to get somebody to open firewall rules for you.

Donnie: Yeah, no shortage of those. Um, yeah, so I think learning from each other is, is a great opportunity of, you know, a conference is a punctuation mark in a good community that has ongoing momentum. That’s a great opportunity for a larger group to come together at one place in time, whether that’s virtual or physical. And, you know, one, one example of this when I think about it is like we’ve got this new experimental tool called Hubtool that’s for like interacting with Docker Hub using a CLI, using APIs in a programmatic way. And so we rolled it out and it’s got stuff like account management, user management, token management, whatever. But what’s really valuable about that is watching how people pick that up and start incorporating that into their pipelines and into serverless jobs and cron jobs. What does it do that makes it into a platform? It’s not about viewing my account, it’s about things like, how do I automate token rotation? How do I make sure that all my images have both the M1 build and the x86 build? How do I monitor whether I’ve used up all my subscription seats or whether I need to go buy more or not? It’s the use cases that are interesting. It’s not Here’s a tool, have fun with it. Because some people will go investigate a tool and dig deep into it and understand everything that’s possible. But a lot of people want some level of guidance of, why would I care? Give me a use case, give me a problem I want to solve, and then help me solve that with the tools at my disposal.

Bridget: [00:33:03] All right. I’m going to tell folks to head over to rusteddevops.com/allthingsdocker when we get this published for the episode show notes, links to the various things people have been talking about. Visit arresteddevops.com/itunes, leave us a review in the iTunes store. Apparently that helps people find the podcast. I don’t know how computers work, that’s confusing to me. But we’re apparently also on Spotify and iHeartRadio if you’re into that sort of thing. And thank you so much to Justin and Donnie for joining us today.

Justin: Thank you for having us.

Donnie: Thank you.

Bridget: This is Arrested DevOps, and remember, there’s always DevOps in the banana stand.