 If folks are going to keep Conchicago, I predict a rise in Generative AI talk. Today, we have with us Grace Winn, Kubernetes 1.28 Release Lead from the University of Waterloo. Grace, it's great to have you on the show. Thank you so much for having me. Excited to be here. Yeah, of course, we are going to specifically talk about this specific release that you've led. But before we go into all of that, just talk a bit about, of course, Kubernetes is in production. Organizations are using it. But of course, as you work with the team, as you work with all the whole ecosystem, because when things become like kind of moving into production, they also kind of become boring. And I do remember when I used to talk to all those folks that they want Kubernetes to be boring. They did not want it to be the shiny, exciting thing. So let's just talk about the state of Kubernetes today that you see. Over the past couple of years, we've seen an explosion of Kubernetes usage across all sort of software orcs. It doesn't depend on size startups use it, big companies use it. And so that creates different challenges. How do we create and release a project used by literally millions of people and make sure that we communicate all the changes that goes into it, right? Because some people use it, folks have feature requests and we have more features coming in. How do we make sure all those features are communicated? How do we know who cares about what? And I think that's exactly the point of a boring Kubernetes project. We each enhancement or feature goes through three stages. And throughout that process, we get in more user feedback and improve until they get to stability and become generally available. Talk about these three stages. For sure, yeah. So the first stage is alpha. So usually alpha and beta could be gated under a feature flag that you have to manually turn on. So you got alpha, you got beta. So this has to be in two different releases. Sometimes there are a couple releases apart just to make sure folks can put in their feedback and as they progress, they get more testing. And then eventually it reaches general availability. GA also call stable. And the interesting things is sometimes we don't need features anymore like a deprecation of Dr. Shim. So that goes into deprecation for a few releases and then we actually remove it from the project and that's called removal. Perfect. So talk about, are there any alpha features in this release that you are excited about or you see a lot of excitement from the community? Yeah, there's definitely one that's stood out a lot. And that is API awareness outside car containers. So folks have been using the sidecar pattern for a long time without having a concrete structure for it. And so with this release, we're introducing an alpha feature. So we're introducing this additional field for the init container called restart policy. And if you set restart policy equals always, that's gonna mean your init container is now a sidecar container. And what that means is that the container is gonna live throughout the lifecycle of your pod. So the KubeBlood instead of waiting for that init container to be completed, it's just gonna wait for it to come up. And there's a lot of implication for this. So one that folks have told me about a lot is service mesh or secret pooling, making sure that secrets are updated as the pod lifecycle exists. One that is particularly interesting to me as a blogging. So making sure that the sidecar pod or a sidecar container is alive before and after your pod. So you make sure you get all that good logging in. Now let's talk about some of the features that are going in GA which don't have any flags and folks can start using them. What are those that you are excited about? Yeah, so one big one for me is non-graceful note shutdown support. So what this means is that in particular situation where your note gets shut down unexpectedly, so that could be hardware failure or you got a broken OS. How do we put in more support so that the pods in that notice not stuck terminating? And so now there is more support to kind of airlift those pods into a healthy note and give your cluster more availability. This is specifically for stateful workload. Anything else? Yeah, one smaller one is expand DNS config. So what this means is more DNS and longer DNS search path. Also, if you talk about this release, if I'm not wrong, each release has its own code and name as well. What's the name for this release? And if you can share the story behind the name and who came up with that name? Yeah, so unfortunately, I had full creative control as lead and so it is somewhere here in the Northern Hemisphere and I'm a big plants fanatic. And so I thought it'd be great to call the new Kubernetes release planter Nettys. But I think it's also a great metaphor for our community. We're so diverse, but each of us plays such an important role in creating this ecosystem. Excellent, that's good to hear. I almost thought that you said plan versus zombies, but not this one. That'd be great for next release. Okay, when you're depreciating a lot of features, then you should call it plan versus zombies. That'd be funny, yeah. Is this release going to be LTS release or I want to understand just how does the LTS work with Kubernetes? Yeah, that's an interesting question. So that is a ongoing conversation right now, not related to the release itself, but LTS is a community conversation and so there's a working group, WG LTS and they started a long time ago, got shut down in 2022 because they achieved their goal and decided on the LTS and then at last cube con, that conversation started again and so now folks are coming together and decided what that number is. Can you talk about the importance of LTS? Also, I want to talk about something which depending on how you look at it, when we look at these open source technologies like Kubernetes, of course you folks keep coming up with new releases, with great cadence, with great education awareness about it, it's not like suddenly this release is being dropped but there was a point where we were seeing a lot of folks were not running the latest release or latest supported release of Kubernetes, they were so big, which actually creates a lot of issues in terms of security, in terms of compatibility, in terms of actually moving from that release to the latest release. So did you still see that problem there? Number one, number two is that can LTS bridge that gap? Because sometimes folks cannot, they don't have all the resources to stay on the latest or the most stable or the most supported version. Yeah, so one thing I'll bring up that is maybe this release is we increase the skew between the version of the control plane and the note version from two to three now. So because we understand that it's easier for you to upgrade your control plane than it is to upgrade your note, now there can be three versions apart and so that's the step in that direction of hoping to better support folks upgrading their Kubernetes version. But the conversation about LTS is an interesting one because obviously vendors and users want to continue to use the version that you're using and be able to upgrade at their own pace. The problem is that we do not have enough resources to maintain 10 versions back. And so it's very much a conversation in the community between balancing what the users want and what we can afford as maintainers. Of course, one of the hottest topics these days is generative AI. Actually, all the interviews that I do, I got four or five a day, generative AI always comes into, and actually a lot of things are moving to production. It's not just a hype, folks are actually leveraging. So sometimes I compare, hey, do you think that generative AI is the same level of technology as Kubernetes or Linux kernel or containers? And they're like, yes, it is. So I'm tempted to ask you that, are you folks also going to leverage some of these generative AI technologies for Kubernetes? Yeah, so there is a conversation going on in SickDocs right now between SickDocs folks and the CNCF about leveraging generative AI to create a documentation, which I think will be a big help. Yeah, code with generative AI, I don't know about that, it might take a while. Yeah, you never know, right? Never know, but I will say, if folks are going to keep contricado, I predict a rise in generative AI talk. Yes, I am looking forward to that event. Now, we talked about Kubernetes, I want to quickly talk about you as well. Talk about what does it like being the release lead of Kubernetes? Yeah, it's a great question. So I'm still finishing university, graduating next year. And so this role has been such an amazing learning opportunity to be able to run such a large, diverse team. We have folks who are industry veterans, they've been doing this for 10, 20 years, their parents, yeah. And then we have folks like me who are students who are newcomers to open source. And so this experience have really taught me a lot about how to run a team, but also seeing everyone coming together literally all over the globe and create these things that the rest of the world can use for free is also really meaningful to me as well. When you work with a large community like that, and honestly, I've been covering open source for a long time, I always believe that open source that you talk about is less about the code, is more about collaboration, is actually more about people. It enables people to collaborate. So talk a bit about when you kind of became the lead, what was your experience interacting with people? Because it does kind of broaden our kind of skillsets to engage with people, especially when we look at open source, they come from different walks of life, they come from different culture, different communities. So talk about that aspect, the culture, the people, the collaboration aspect. Yeah, I think that's the part I enjoy the most is that no matter where people come from, no matter who they work for, we all have this one common goal. And so I've been working on the release team for over two years now. And I think that gave me an understanding of the community and from that I was able to know who I can pull into a certain situation. Obviously we have our own challenges, for example, folks are in different time zones. We have a lead in Australia and he's not able to answer us as fast as the folks in the Northern hemisphere. And so those are things that we're always trying to improve and communicate. So right now for the release team, we have two meetings a week and that's like a normal release team meeting and we have one for like APAC post meeting. Grace, thank you so much for taking time out today and talk about this release, of course. I loved about the Kubernetes part, but I also love the culture part that you shared because I think working with open source communities actually enhances us, improves us, how to interact with folks. We don't talk much about that, but I feel that is the people is one of the biggest part of being part of any open source. So thanks for sharing those. And I would love to chat with you again, but I really appreciated the time today. Thank you. Thank you for having me.