 So we do have some ODF product managers here, and I've invited them on stage, even though they're not red hat and engineers, speakers, if you want to come up on stage and if anybody has any questions, we'll have Lenka running around and so appreciative to her for, come on, you're just leaving me up here by myself. Well, you would, you would. I mean, let's talk APIs, I'm just kidding. Okay. Thank you for joining. Okay. I know, Peter. Oh, come on. I know. Let's see. Thank you. Then we'll do a quick round of introductions. If you were a speaker, please join us. I know people had questions for you, and I was so rude and I shut them down so we could keep going. We're almost to the reception. Fabian, I know. Right? Okay. I'm going to pass this around really quick. Now, also, memorize these faces. So as you go through KubeCon, if you want to have some deeper discussions, deep dives, you know who to reach out to. So I'm going to start over here with Michael. Hello, Michael Rivnak. Been in containers for a while. Installation and bare metal are maybe my things. Hey, everyone. My name is Peter Hunt. I mainly work on cryo and sig node and KubeLit things. So anything pod running related? Hey, everybody, Michael McKeon. I work on cloud infrastructure, cluster API, cloud controller managers, that kind of stuff, auto scaling. Hi, my name is Daniel. I'm standing at the edge, which makes a good image because I'm the product manager, OpenShift Edge. That means I cover single node OpenShift, multi node OpenShift, remote worker nodes, if deployed to edge locations and most prominently, micro shift. Yeah, I don't care about it, I'll be a mess. I'm Chris. I'm one of the ODF PMs. And I'm not going to be at KubeCon, so just hit me up now. Hi. I'm David Eads. I currently work as a staff engineer for OpenShift. I spend a lot of my time upstream in Kubernetes for the past eight years or so, probably specialized control plane, auth, etc, integration with OpenShift at sort of whatever level we need. Hello, I'm Augustin. I'm a principal engineer from MAMadeus. I'm currently focusing around the prometo UI, how to improve it. I also created Perces, just repo, not all the code. And yeah, I'm also working on that currently. And we are in close collaboration with Reda to improve it. Hello, my name is Alban from MAMadeus as well. And I'm, let's say, expert on prometo and Thanos topics. Hi, my name is Ricardo Noriega. I'm working on edge computing related topics and micro shift. And hello, I'm Michelangelo Ajo. I also work at edge computing topics in Red Hat. I have a past in hardware development and I think that works really well with edge and I'm trying to make things move there. Karina Angel, OpenShift Foundations and OpenShift Commons. Hello everyone, I'm Vanessa Martini. I am the PM for OpenShift Analytics and Observability UI. So I'm one of the PMs of OpenShift Observability. Hello everyone, this is Naina. I'm the OpenShift Serverless PM and Serverless has Knitif serving for event-driven functions and Serverless logic for workflow orchestration. And you can come and talk to me about other layered products like service mesh, GitOps, pipelines. Hey, I'm Damiano Donati. I am an engineer on the OpenShift class-stream infrastructure team, MAPI, cluster API, control play machine sets, all those things. I'm Joel, I'm the team lead for the class-stream infrastructure team. So all the same stuff as Dam and Mike. I'm Fabian. I'm taking care of Qvert at Red Hat. Yeah, you just heard who I am. Norris, Samo Sanganko, SVA and yeah. Ask me questions if you have any. Peter Golovski, IBM. I'm responsible for architecture overall, the IBM products on running on OpenShift. Any questions? Ask me. Okay, now's your chance and I see Thomas didn't come up here but I'm not gonna call you out. All right, first question. I know you weren't shy earlier. We are really close to the reception though. Right? Okay, what about up here? Are there any questions that you'd like to ask of the audience? I know. Yeah, absolutely. Can I get a raise of hands? How many people here have been cluster admins on something like say at least five node clusters? That is a lot better than I thought. How about total number of clusters? Cluster admins for more than 15 clusters. Oh, still a few. 50? Five zero? All right, still got a few. That's very interesting to me and how many people would call themselves platform extension developers. So someone who's gonna write a CRD, maybe a controller to store some data. All right, look at that. Glad to see all that. If you have questions about that sort of thing or just wanna talk about it, I am very interested afterward and I'll be at the reception. Thank you, David. Also ask him about PKI. Hey, one presentations earlier talked about hyper shift and there's been a lot of talk about how control planes are evolving. So would anybody up there like to talk about that? So hyper shift is being developed by a sister team to ours which I think is why I've taken the mic on this. I know it's coming later in the year for basically a self serve. You can have the dense control planes running as pods on a single cluster and then you just have your workers spinning up to it. We're looking at getting that into Rosa, into ARO and the idea is we save money, right? Like control planes can spin up in minutes rather than the 30 or 40 minutes it takes for an IPI installed today. And yeah, you get much more dense compute custody. They've, yeah, I don't know what you wanna know about hyper shift. I can certainly give you the contact details of the lead and then if you wanna know more I'm sure they're happy to talk. Thanks Joel. All right, any questions? All right. Hi, so specific to like observability I guess. You'd mentioned an observability operator at a pretty high level. Could you talk about any more details for like in this GitHub links? So at the moment it's still working progress but if you get by the Red Hat booth in the next days I can give you a better overview. But definitely we are really putting a lot of focus on that, on the simplification of the data collection part and how you can adjust what you want to collect. So if you will come by I can show you some more details definitely. There's hopefully an easy one for David. You mentioned that the choice of cell was around one of the advantages there was that if a calculation is too expensive you just won't run it. Will that be surfaced as an error? Is there a way, what does that look like? If I have a validation that is that complex what are my roots out of that problem? Right, so we're still looking at that. Admission is at an alpha level so I guess I should back up one level. The first set that I presented was static. It exists inside of a schema in our CRD and those realistically aren't gonna be very complicated in terms of runtime unless you do some really unusual things. And in that case your solution is to look at your prometheus metrics you're gonna see that it's expensive you will get a rejection from the API server and you'll be able to go and say this has to be fixed and if you own the platform extension I can't remember if you raised your hand but if you own the platform extension and you hear that you were too expensive you need to fix it because it's everywhere, right? And then we have another part that runs in admission and I think cluster admins are probably aware of that because they've had webhooks go bad and there were a bunch of you at least some of you have had a webhook go bad at least once. And when that happens you get a runtime error as well and we're still looking at the API for that. Do we wanna have an API that says I'm willing to budget this one higher? Maybe. Do we wanna have a flag that says nothing more than this? Maybe. But right now it is set up to give you a runtime error you will be able to see where it comes from and then you'll be able to see it in your metrics and make a decision am I gonna change it or not? There's a question over here Lenka. Thank you. Hi, my name is Felix. I'm more kind of cloud native at the big cloud providers Kubernetes at the moment with my team and we are switching over and searching for new distributions for on-premise operations. So my question is, how would you sell me OpenShift as the best on-premise distribution? Do we have marketing up here? So I would look and say that the roots of Red Hat are on-prem when you compare against other vendors in the area that's actually really valuable. The other thing that we have is we have it all the way down to the OS. So if there's a problem and it's on-prem, we have it. We have worked with people to bring new hardware up to date and say, okay, you need this hardware enabled we can help make that happen. And our linkage to that hardware, the accessibility that we give it to our clusters is really, really good. Whether you're on bare metal whether you're on VMs it works extremely well and it was developed to serve there as opposed to you have an offering that runs on a cloud and now you're trying to fit it in. There could be differences there, right? Is that satisfactory? All right. Oh, sure. I cannot do that because it extends to the whole platform. For example, it starts with AirGap disconnected installations. Have you ever tried to install EKS AirGap? Good luck with that. And we have it just from the runtime platform but also for the management development tooling if you need to develop your solution completely AirGap and we have customers needing to do that all the management aspect if you think about for example advanced cluster management, Quay, ACS all of these can run completely AirGap on-prem but you can also of course connect them to the cloud if needed. If you need to add something just for a couple of weeks more compute power add a cluster in Azure AWS, whatever and add and scale out as needed and that is the beauty that you can mix and match as you need. So on the technical side, as you mentioned also for example on the AirGap side there are different integrations like with the Cloud Control Manager also where you can directly spin up worker machines from vSphere so you see them in your vSphere topology and this developer experience I call it always extends towards the UI so I don't think there is like a holistic offering where you currently have like the UI aspect that you have on the cloud where you can see your applications you can organize them and you also have the technical part and the good thing is usually it's based on open standards so that's also gives you a benefit to use your existing knowledge like Istio or Canada for example or like obviously the GitOps and so on so yeah, that's what I want to add. This one is not marketing at all. Are there any other questions? I would like to follow up on that question on how many OpenShift clusters do you administrate and run? Being the Edge product manager I'm obviously interested in do you have clusters in general? Let's say Kubernetes clusters doesn't have to be OpenShift could be something else don't be shy on that. Do you have clusters at the Edge? Edge being defined as not in your traditional data center not in your cloud somehow restricted in resources like networking, physical size be as generous as you think. So if you have something running at the Edge Kubernetes-based raise your hand please. Sorry, what was that? Yeah, of course, yeah, you're a smart home. Okay, thank you. Thank you.