 Hi. Thank you, everyone. So thanks for making it in such numerous numbers to my lightning talk. What made your container fat? So a few words first about me. I'm a software developer working at SUSE, part of the developer engagement program. I also do other stuff, mostly in developer tools, testing. I'm also a member of the Fedora community. And in case you are not totally bored after this talk, you can also stalk me on some social media where I occasionally write maybe interesting stuff. You can be the judge of that. But let's get to the actual topic at hand. So we at SUSE, we do all kinds of stuff. And one of them is building containers. And if you start building containers, you occasionally run into the issue, well, gosh, the image got bigger. Oh, well, where? And then you start unpacking container images and you start analyzing them. And it starts to get confusing pretty easily, pretty fast. So where is actually the size increase coming from? So maybe you want to take a closer look. Then you figure out, hmm, okay, I have a big container image and got 20 layers. Which layer is now the actual culprit? Where did it get the big? And there you can't really go into the old trick that you just launch the container image. You install NCDU and you take a look. Because then you have to find the actual layer. So you can unpack it and you put it somewhere in your TempFS and you get super. And after you've done it 50 times, you have a huge temporary directory with all kinds of layers extracted. And you forgot which one is the actual one that you were looking for. So then, so there's fortunately a pretty cool tool for that. It's called Dive. But maybe you like prettier graphs. And I like the GNOME Disk Usage Analyzer, which produces pretty nice sunburst charts. And I wanted something like that for container images as well. And also, what would be really nice being able to compare images side by side. Since I didn't see something for that, I wrote something. So let me show you this tool that essentially shows you a sunburst graph of a container layer. So if you want to see what actually is big in your container image on the file system, you can do it with that. So let's jump into a short demo what the, how the tool looks like. This is a real, it's just a web app. So you can open it in your web browser, you give it a container image. It will show you the various platforms via which you can pull it. So you can select any architecture that you'd like, you tell it to pull it down. In this short demo, I already had it prefetched so that we don't have to wait for five minutes because my internet connection sucks. Then it will start analyzing it. You'll get a table of all the layers that are there. And then you can plot them. You'll get a nice sunburst graph. You can click on them. It will zoom in. The over, you'll see as one of the overlays is how big the actual file is or the folder. You can select another layer. And essentially it will give you all the interesting information per layer. So that should be that one. And now then I thought, well, okay, so now I can see what each layer looks like. But now my image evolves over time. And so, well, something changed. What? So for that, it can do comparison of various layers and will essentially show you the same picture. Just side by side. You give it two layers. And everything that's red means it's bigger. Everything that's green means it's smaller. Blue means it's the same. And blue, which, well, let's just run the demo. So if you could just click on any of that, it will show you the same folder in the corresponding image. And if you take a look at, if you at some point see something, yeah, there should have been a blue one as well. That means it's just present in this image. So since I'm running out of time, let's just take a look at a few of the features that are in there. You can take container images from a registry or from a local archive. So you don't have to even push it anywhere. You can analyze images from various architectures, so you're not limited to your current systems architecture, which also became crucial at some point. You can store them in a database, currently that's SQLite. It runs rootless by default, if you like that. And if you want, you can run it on your local system as a containerized web app. And if you don't want to type those commands, you can also just give it a try yourself. And with that, I'd like to thank you for your attention.