Cloud Native Computing Foundation KubeCon + CloudNativeCon North America 2022, 12 Nov 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Keynote: What a RUSH! Let’s Deploy Straight to Production! - Whitney Lee & Mauricio Salatino

Description

Don’t miss out! Join us at our upcoming event: KubeCon + CloudNativeCon Europe in Amsterdam, The Netherlands from April 17-21, 2023. Learn more at https://kubecon.io The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.

Keynote: What a RUSH! Let’s Deploy Straight to Production! - Whitney Lee, Staff Technical Advocate & Mauricio Salatino, Staff Engineer, VMware

Speakers: Mauricio Salatino, Whitney Lee
Feel the adrenaline of deploying to production during business hours! Join as Whitney and Mauricio discuss how to empower application teams to be happily productive, and to enable them continuously deliver value to end users. Learn how a platform built on top of Kubernetes can facilitate the creation and consumption of complex infrastructure, so that developers can have the tools that they need at the moment that they need them. By understanding best practices around platform engineering and how a platform team can provide a beautiful developer experience, you can pave the path to production for the entire organization.

Watch a live demo showing tools including CNCF projects Crossplane, Helm, VCluster, Knative, and ArgoCD to gain a practical understanding of how to adopt, integrate and glue different tools together to build a platform that your application development teams will love.

A

Hello, we're so happy to be here what a gift! Yes,.

B

We are here to have some fun, so are you here to have some fun? Let's have some come.

A

A

Imagine you're the developer in this diagram and you work at a company that provides a very important software product rainbows as a service. Now you're, probably your company is dominating the rainbows industry and it wants to stay on top. So with commissioned Gartner reports, we put together a think tank of all the industry, thought leaders and we've identified the next big Trend in the rainbows industry. Do you want to know what it is because I.

B

Really want to know what it is with me tell us it's.

A

Spiders, what what I think spiders are literally the opposite of a rainbow so anyway you're the developer in this diagram and it's your job now to add the spider to the rainbows and push it to production, and when you do that, your change is immediately live in front of all of your companies and users. This is scary and I'm, not just talking about the creepy spider here.

A

This is scary, because, if you slow down, if you, if your application has problems, you can slow it down and cause problems for the end users or worse, you could be the cause of your company's most dreaded fear down time. These are high stakes. Well,.

B

Well, why causing downtime or frustrating your users is something that you should try to avoid. We need to agree that this diagram here is a drastic oversimplification of what happens when a developer pushes gold, okay,.

A

Okay, I'll give you that it's another, it's an application! How about this? That's a little more accurate. Wouldn't you say ah it looks.

B

A little better, but what is the developer there? Oh.

A

Oh okay, uh the developers way over there on the left, the production's on the right yep. But don't don't worry if you don't understand everything that's going on in this diagram? In fact, don't worry if you don't understand most of what's going on in this diagram.

A

The point is for us to show you that it is a lot. It is complicated and it has to be complicated when you're building complex distributed systems that are meant to scale across millions of users.

B

That's true and the question that we are asking and trying to answer during this presentation is how much of this entire landscape do. You need to actually understand in order to be efficient at your work. Yes,.

A

So to begin to answer that question: let's talk about kubernetes.

B

By by using kubernetes and the power of the kubernetes, apis companies have a clear separation between infrastructure and application development teams. uh Infrastructure teams can focus on provisioning hardware and VMS if they are working on premise or choosing their favorite cloud provider. As most of this, Cloud providers already support the kubernetes apis. On the other hand, application development teams have a very powerful API and declarative API to deploy their workloads, no matter where that infrastructure lives.

B

So that's a good thing, but we need to agree that this is yet again another drastic oversimplification of what happens well.

A

It's very helpful that application teams now don't have to deal with the infrastructure directly. There are still some challenges for one: it can be a challenge to train big teams to use kubernetes and, let's be honest, not every developer wants to learn how to use kubernetes. Another challenge is that there are so many amazing projects in the cncf landscape and it often falls on developers and application teams to research these projects and figure out how best to extend their kubernetes cluster for their particular use case, and our application team needs to be writing code.

A

They should be thinking about adding the spider to the rainbow's application. Why is the.

B

Spiders, so welcome the heroes of our story platform themes, more larger and mature organizations that have been have been working with kubernetes for so long time are starting to create platform teams to build platforms on top of kubernetes.

B

They do that by trying to understand what those application development teams need and try to hide all the complexity of installing managing and operating these tools, putting everything behind a very friendlier platform API this platform API, can be used by application development teams to self-service, to have access to all these tools and things that they need to work using a self-service approach. Look at that platform that.

A

What are friends very friendly, friendly, API I want to invite them to my birthday party. You should you should I hope they say. Yes, okay, so the platform team starts to build the platform first by collaborating with application teams. This is a very important step. They need to deeply understand the tools and processes that the application teams are currently using and what those pain points are.

A

Then they can figure out the best way to help and encode that knowledge in to our friendly friendly platform, API that then, the application teams can access in a self-service way. So, for example, our application team needs a development environment to be able to add the spider to the rainbow's application.

A

So our developer asked the friendly friendly platform API for that environment and lo and behold, without any human intervention that environment gets created and it has a speedy spider cache and the spider database, and not only that's but the environment comes pre-configured with security and governance policies that the application team doesn't even need to worry about. There.

B

You go, and this takes us to our first demo of the presentation where we are going to interact with our platform apis to request a new development environment. So what we are seeing here in this recorded demo I would I would love to rely, but I couldn't so basically I'm part of the arachnid team, and what we're doing here is we're going to request a new environment, and we can do so by sending a request like this, which basically we have extended the kubernetes API to understand about environments.

B

You can select the type of environment by setting a label and also then, as you can see there, setting up some parameters that are very, very specific to the the the needs that I have right, like in this case, having a spider database or a cache. As you can see, I'm sending like this request using cubectl, because it's a kubernetes resource and voila the platform just created an environment. For me, you can actually run this limo live. We will share some URLs later on, uh because it's a kubernetes resource.

B

You can query all your environments using qctl or any other tool in the kubernetes space, and then you can see there that the environment is not ready yet, but we will use this environment in our next demo and we will deploy some workloads into that yep.

A

Excellent, so what we just saw was Mauricio interact with the friendly platform API and behind the scenes head platform, API used cross, plane and Helm together to create that development environment. That development environment is an isolated virtual cluster made with fee cluster, and it has k-native serving pre-installed into it.

B

The good thing about these things is that you don't really need to understand how cross-plane Helm, canadi or B cluster works, because that's the whole purpose of having a platform, a Platform Theme, creating these platform apis. That allows us to send a simple request: to create a bunch of things for us.

A

Excellent, so we saw how the platform team can help application teams provision development environments in a self-service way, but platform teams can also make the developer experience better by providing more widely scoped tools that, for the application, team makes the right thing to do the easy thing to do so. For example, they could create templates and quick starts so that developers can start their projects easily and by following company policies right from the beginning.

A

They can also curate, maintain and continuously patch tools, like Cloud native, build packs that automate the way application source code gets built into a container images. And finally, the platform can provide a catalog of ready-made solutions that application teams can then extend so Solutions, like tecton pipelines or Helm charts and.

B

That takes us to our second level right, we're going to basically create and deploy a function using a project that is called k-native functions into the development environment that we just created. So we are. What we are going to see here in the demo is that we are going to first check that our environment is ready to be used. That's why we run that command. It is ready, and now we are going to use big cluster connect because we use big cluster to create the cluster.

B

So we need to connect to it, and now we are in our environment. We can use it, we can do whatever we want in it and what we're going to do is we are going to create a new function again using this K native function functions project. For that we will go into separate directory. We create our function directory called spiderize, because we want to bring some spiders into our rainbows, and then we use this command func create pointing to a repository that contains some templates.

B

In this case, we are using a template that it's called spiders and that allows us to bring all the things that we need to draw spiders on top of our rainbows. Once the function is created in this case, using the go programming language, we can change the code, we can modify it any way we want, but we don't have time for that now, so we are just going to deploy it using fund deploy when we run Funk deploy.

B

This is using again this project that is called Cloud native buildbox to transform the source code of the function into a container image. So the developer doesn't need to worry about how that transformation happens and it gives the platform team the ability to create how that process works by the end of this command. We are going to get a URL down there in the in the screen, where we're going to click and we're going to see our function deployed in our environment.

B

So if we click into that link, we should be able to see that now we have a spider and a spider on top of our rainbow in our development environment. So.

A

Trendy, it's pretty pretty I, feel so cool right now, so what we just saw is Mauricio connecting to the development environment with v cluster connect command and that environment already has k-native installed into it.

A

So Mauricio could immediately use the funk CLI to create and deploy an application without any Docker file or any yaml, and the platform and platform team helped enable this sweet sweet developer experience by not only providing that development environment, but by providing templates to get started in a quick and easy way and then providing build packs to automate the build of source code to container image in a company happy way there. You.

B

A

Okay, so so a couple of ways now that the platform and platform team can improve the developer experience and the development life cycle. But now how can the platform team help that finished application find its way to production so.

B

You mean that we cannot push straight.

A

To production, I'm, sorry Mauricio developers under no circumstances should be pushing streets of production. Instead we're going to rely again on our heroes, the platform and platform team, this time they're going to provide tools and best practices to help that application find its way to production. So in this example, we have a get Ops approach to do that. So we have the production environment configuration stored in a git repo and we're going to use a git Ops tool like Argo, CD and Argo.

A

Cd will watch that git repo and if any changes are made, they're going to sink that change to the production environment. So then, when our application team is ready for the spiderized application to see the spotlight, they submit a pull request and then automated testing happens and when all those tests are green, the pull request is merged. Argo sees the change in the git repository and then the application is deployed there.

B

You go so, let's see that in action and for that I will walk to my computer to try to do something. It will never work right. So we are going to take our function into our production environment on our production environment is actually rainbow social service.com.

B

So if you go to that website, it's actually working there and we have configured this production environment using a etops approach, as we mentioned, and we have this repository called kubecon production in my GitHub account where we are configuring, the application using K native serving here, that's why we have a single demo just to configure the entire thing. So if we want to apply changes to this application again, as we mentioned, we will have a pull request with those changes. In this case, we want to add the spiders modifications that we just created.

B

um This is where the platform kicks off right like this is where we need all the validations to make sure that these changes will actually run in our production environment, so performance test integration tests. All the things that you need to do will be validated here when this is ready, we can actually merge these pull requests to apply these changes into our live environment. So that's exactly what I'm doing right now.

B

So if you access rainbow service a video now you can access again uh us because, as soon as these changes are done- and these changes are synced, we should be able to. You know, see all the changes applied into production and if it's not true, I should be able to refresh here in my production environment there in the screen and there we go. We took all the spider from development to our production environment, hey.

B

A

B

A

So bleeding edge rainbow technology.

B

A

Spider's rainbow tag on Twitter is trending so hard right now there.

B

A

Go all right so, to recap the whole Demo First. We use that friendly friendly platform API to a self-service provision, a development environment then because candida was already pre-installed into it. We use the K data, Funk CLI command to get that spider the create and deploy the spiderize application with no Docker file and no yaml. Then the platform and platform team helped our application find its way to production in a safe and automated way.

B

So Whitney, what do we want people to take away from this beautiful presentation.

A

The big thing to remember from this application is that platform team zest to manage all the complexity of getting an application into production, and this frees up application teams to do what they do best, which is to write code so specifically, platform teams do this by collaborating with the application team to understand what they need and what their pain points are, and then to to add that knowledge into the platform API, where the application team can access it in a self-service way. There you go.

A

The platform team also improves developer experience by encoding things like best practices, architectural patterns, security and governance policies into the platform team into the platform itself, so that the application team can actually doesn't even need to worry about it.

A

And, finally, the platform and platform team helped to automate the path to production in a safe way. So.

B

The bottom line here is platforms and platform teams are there to make application development teams happier and more efficient, and they will take care of taking all the changes and all the cool features that they build safely into our production environment. So one more thing before we go there is that GitHub URL there, where you can actually run all these examples in your laptops or in your kubernetes clusters. I strongly recommend you to play with this Technologies and see how this works and that's what we got. Okay,.

A

B