 So, welcome to DEF-Room. We have the third monitoring DEF-Room in a row. We've had the largest or second largest room every single year, and I recused myself from all DEF-Room assignments for obvious reasons. Still, we managed to completely fill up this room several times during the day, so this is just awesome and shows how much interest in that topic exists. And also it goes to show that if you're interested in, like, say, like, 70% of those talks, chances are you do not want to leave, else you might not get a space anymore. But that's your choice. So, I'd like to especially thank Ben Kochi, who is not yet here, Carl, who is also not here, and Fabian Steder, who is making sure that the door doesn't bang all the time. You will be seeing us a few times around, and seeing, and doing stuff. So, let's quickly go through the content. Of course, I know we have this online. I know it just broke because we had to switch out a speaker last minute. So, Penta Bar broke and I didn't manage to fix it yet. Anyway, we'll start with Rosario, who will be talking about Savix. Savix is kind of, or hopefully we'll start with him. But in my mind, this is, like, the best last-gen monitoring system on Earth, and it literally kept my sanity for almost a decade. If I hadn't had Savix, I would probably be in jail or something. Then we have Simone Minority. I hope I get those names right. With augmented network visibility and high resolution matrix. So, we're talking, like, second metric, second resolution. So, we're not, or he's not delving into microbursts, but into bursts. So, and this is also already really, really useful to just see, do you have, like, the full pipe for half a second or maybe two seconds? Of course, this can already massively hurt your performance of pretty much any application. And it at least gives you some hopefully way to detect if you had microbursts, because those are even bursts, especially when you have any storage system. Then we have JBD, Yana Dogen, critical path analysis. I'd like to say that there's only very few ways to actually think about what observability is and what matters in monitoring. And going through the critical path of your applications and of your user stories of what your users are actually trying to do is one of the best ones. Of course, it identifies all those really, really important bits and pieces, which might not be immediately obvious, only obvious. There's Carl. He also gets a thank you. Then we have myself. I'm standing in for Priyanka Sharma. Of course, she had visa issues, which just goes to show that visa sucks and diversity also has some unseen challenges to it, or unseen to at least anyone who's white and European passport and no problem with all the things, but others have issues. So, yeah, that kind of sucks. She couldn't make it, so I'll be having a full talk. Initially, it was planned for me to open the staff room with just like a short version of this talk. Now I'm doing the full one. Then we have Tom Wookie, who's going to talk about Loki, which basically takes the metadata system and the label system off Prometheus and uses it for logs, which is not full text search. Like, for example, you would have Splunk or Elk Stack, but on the other hand, it scales way better and it costs way less. So it's a good point to be in between, and you should definitely have a look at that one. Then we have someone with a really good surname. It's the same as mine. So if you want to do latencies, is the classical like 99, 95, is this the correct thing to do? Is it always the correct thing to do? When do you want to do something differently? And all these kinds of considerations come into play. Then we've got Rob Skillington, who works at Uber, so he did Uber for Metrics. M3DB is their internal stuff, which they've been working on for a few years now. And it's quite interesting, to be honest. It's similar to all the other current-gen metric things, but it's pretty interesting in and out of itself. Then we have Ion Lermont. I hope that is somewhat correct, at least. Talking about how to do monitoring for Tor without exposing users to surveillance. Obviously, you want to avoid identifying users on Tor networks. On the other hand, especially as people are anonymous, you need to have good visibility into how the system runs. And this trade-off, that's what he's going to talk about. Then we have Peter Seitzeth, who's going to show us the EBPF exporter of Cloudflare and also talk a little bit about what EBPF does and how to get new information out of your Linux systems. Of course, this is kind of a game-changer within Linux. You just get, in some places, a lot deeper information than you could get before and a lot more consistent. Then we've got Bartek. He will be talking about Thanos, which basically takes normal Prometheus, which normally is a single binary. It runs as one single thing, and basically it takes a storage and it puts it out into object storage, which allows you to do all kinds of neat things, and he'll be talking about that one. Then we've got Metalmanze and Sergius Ubonic talking about Kubernetes and how to do monitoring of Kubernetes with Prometheus. For those who don't know, Kubernetes and Prometheus wouldn't be as big if the other one didn't exist. Of course, they are in a very, very tight interdependency. I mean, Prometheus doesn't really depend on Kubernetes, but the rush of Kubernetes propelled Prometheus forward, but you couldn't really be running Kubernetes unless you had some monitoring which does labels. This is why they're really closely coupled, and those two are really, really good at what they do, so we'll be seeing a lot of that stuff. Once more, if the next speaker is here, you can come up and prepare. Of course, we would like to switch over in two minutes. Yes, exactly two minutes. Then we have something about James Schuben about trying to take your monitoring data and act on this automatically, so get nearer to this holy goal of self-healing systems. Then Hannah Suarez about why it makes sense to have centralized logs and not just have bits and pieces and silos of your data and small islands of your data and then have fun with trying to discern anything, especially in incident response. That's the second class. We have Ilya Agenov on writing asynchronous SNMP agents. SNMP, for those who don't know, is pure distilled pain. It sucks, and you still have to use it. Then we have monitoring for Internet of Things and how you can tie MQTT and all those things into a monitoring story. Finally, we have Michael Sturda how to do LDAP monitoring. There's a few things I'd like to say. If you see trash, don't care if it's yours. Take it with you. Take it outside. Don't put it into these. Put it back in those trash cans, not those, because those are always full. Please take it to those trash cans. And also, please respect the speakers. They're doing this mostly their own free time. Some even travel on their own money to come here and speak. We want to maximize, well, Q&A time, not Q time, but whatever. So, and for us to maximize Q&A, we really need you to be quiet. You can enter, you can leave, no problem, but be quiet. If you're not quiet, we will shout you down. This is not because we're evil. This is because it's about respect for the people who invest their time to do this. Thank you very much, and if we have the speaker, then I'd like him to come up, and if not, we have an issue.