 My name is Alejandro Pedraza, I'm a software engineer at Boyant. Linkerd is the fastest and lightest service mesh. This app allows you to vote for emojis. You have a leaderboard to see what are the favorite emojis among your users. We have deployed these in a cluster whose pods are listed here. We have under the emoji voter namespace, the components that make up our app. We also have the Linkerd control plan installed and Linkerd viz, which is an extension that gives you extra visibility that we will use in a moment. We have ambassadors, traffic manager pod, which coordinates all the intercepts happening in this cluster. Now, let's say users have been complaining about problems in our app, even though it's working fine for us, as we just saw. In order to take a deeper look into what's going on, we're going to launch linkerdviz dashboard. The first thing we see is that the emoji voter namespace is not operating at a 100% success rate. Let's click there. We see the components that make up our app, emoji and votebot are operating at a 100% success rate, but visibly voting and web are the ones having trouble. Let's click on web. In this diagram, we see how traffic flows into and out of web. We see that web depends on a service called voting, and apparently that's the service responsible for the errors. So let's click there. If we scroll down a little bit, we see all the requests coming into this pod. Everything is looking fine so far. But we start seeing problems. Vote. Donut is failing consistently, so most likely that is the source of all our problems. Up until now, we have used linkerdviz to identify the location of the problems in our cluster. Now I'm going to hand it over to Edidion to show us how we can use telepresence to debug and actually fix the problem. So we're going to go ahead and create what we call an intercept, which is going to be routed to traffic intended for the voting service in our local cluster, to the voting service in our local mission. As you can see in the command, it also has a special KCC header linked to debug and what that is going to do is basically signifying that only request with this header should be passed to the voting service on our local mission, and any request without this header is going to be a voting service on our Kubernetes cluster. So now we have that working. And the reason why we're doing that is to ensure that while we're testing and debugging, it doesn't affect the development of our team. So with that, we're going to quickly start the emotional service on our local mission, and then I'm going to pass the header, which is debug and keep going. You don't have to do this. This is more of a manual approach. But if you've logged in to Baccalaureate Cloud, you'll automatically create this personal intercept for your emotional provider. It could be a route that you can share with your teammates. So with that, if you vote for different emotional services that it works well, but if you vote for the general thing when you come up next, you'll see that it's not going to work well. And that pretty much compares to what I had to have located within the intended. So now we know what the problem is. And the next step is to actually go to our code base and fix that issue. So I'm going to open up the API. The profile where the Donut function is in. And with that, I've scrolled down to the Donut function. And you can see there that it has a couple of code, but then the section there says Valencia Pirate app is the best example, which is not true. I tried this, and it's fantastic. So all we need to do is change that and false to true. So I'll quickly do that. I want to change that to true. We're going to save the changes, go back to our CLI, and that's the service again. So if what we've done, I think what you know, we see that it's going to work well. And that concludes our being able to test the local services, all that we've got services in our cluster. And we're going to then produce the production. I believe that is working well. That's it. That was fast and easy. Not just because this is Lightning Talk, but because these tools were built with usability and as the first and foremost concern. You can reach out to us after the session. Thank you.