keptn Keptn Tutorials, 10 May 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Using ArgoCD with Keptn: Developer Experience in 10 minutes

Description

Thomas Schuetz, the Keptn maintainer, goes through a use-case of running Keptn and ArgoCD together. In this presentation you will learn how Keptn can orchestrate Argo deployments, how to do continuous deployment with Argo while validating your apps with Argo, and other possible integrations through webhooks and notifications. Thomas also speaks about doing promotions via pull requests which is currently in progress.

Learn more: https://keptn.sh
Get started with tutorials: https://tutorials.keptn.sh
Join us in Slack: https://slack.keptn.sh
Star us on Github: https://github.com/keptn/keptn
Follow us on Twitter: https://twitter.com/keptnProject
Sign up to our newsletter: https://bit.ly/KeptnNews

A

Hello, everyone and welcome to this captain webinar about using argo cd with captain. My name is thomas schutz, I'm a principal engineer at banner trace and the captain maintainer. So today you will learn why it's a good idea to use captain with argo cd and how this integration looks like so, first things. First, why use argo cd? So I think we all love gitobs and the possibility to have our deployment code versioned in a git repository and argo is very good in doing githubs and we get continuously reconciled deployments when you using iocd.

A

Furthermore, we get a nice visualization over the current deployment state, but also about the things which happening behind the scenes in kubernetes.

A

Yes, furthermore, we get some change in drift detection, so if some objects changing humanities um and on no more corresponding to the to our intent on the git repository. Argo can detect this and can also take actions upon this. And last but not least, argo is able to deploy a lot of things, so we can deploy. Kubernetes manifests as well as salem charts, but we can also use customize so, secondly, why should we use captain?

A

So captain is very good in taking decisions based on based on observability systems, so um captain is in a slow driven application. Lifecycle, orchestration tour. Furthermore, you can bring your own tooling or write your own integrations. So um when you have some functionality in your current deployment pipeline, you can use this as before. Simply by implementing this in the job execution service well write your own service. If you got some kind of integration which is not there at the moment.

A

Furthermore, captain can orchestrate your deployment across multiple stages, and last but not least, you can integrate create a lot of sli providers in your workflows, such as dynatrace, prometheus and also data talk.

A

And yes, what you will see today is that we will get a deployment repository and argo is able to detect changes on this on this repository and when argo finds some changes in this repository, it will notify captain in our case- and this is the kickoff sequence here- and in this kickoff sequence, we can decide if we want to apply this, this change or not. uh Yes, we will apply it and afterwards, captain will deploy exactly this. One what's happened.

A

What's happening in the in the repository uh and argo will apply this last but not least, um argo will notify captain about the state and that it's that it has finished its work and afterwards captain will proceed in running tests and also doing evaluations and doing the promotion step.

A

So I think we talked enough and let's proceed to the demonstration.

A

So at first here we go. We have our cloud automation. Instance. We see that no sequence is running at the moment and we see here our deployment um manifest. So we see here typically our application, which has some specialities in there. So we see that we label this. This argo application with some additional labels like the stage like the captain service and the project.

A

Furthermore, we see the chart and the repository url of this of this um application and also see the target revision of all of this, and the first thing we will do today is we want to. We want to simply deliver a new version of this application, so we'll simply will only change the target revision here in the manifest.

A

And let's commit these changes and what.

A

We will see in argo in a few seconds is that the application we specified here so the potato head in the stage death gets out of sync, but we also see that um the target revision of this application is the new version of the of uh which we defined before, and another thing we see now is that we get a new sequence triggered in our cloud automation instance, and therefore we see that our approval step is pending and we will simply approve this now and if we get back to argo, we should see in a few seconds that it gets thinking that it will sink, and you see that the sink is progressing at the moment and it's it's healthy now and when this is healthy, we also see that our artifact delivery step has finished.

A

And now there are some test running and evaluations will be triggered in the. Meanwhile. I can show you a few things: a few nice things about this integration.

A

So in captain you can click on this argo cp button and you will get directly to this argo city screen where you see the current deployment state and furthermore, you can also have a link to um you also get a link to your application here.

A

So this is our test application and we see that the version 0 2 0 is deployed.

A

So, as I said before, currently there are tests running and now the evaluation gets triggered. So we are talking to our slide provider at the moment this is dyna trees and we'll get some results in a few seconds in the meanwhile, I will get back to the presentation and show you what I did before before this demonstration.

A

So behind the scenes before us, I started all of this. I set up the our applications. I created the captain project, but also the service incap.

A

Furthermore, I set up some notifications in argo, so, firstly, when the application gets out of sync, it will notify captain that something has changed um and furthermore, when the deployment of the application gets finished it will, it will also notify captain about the finished deployment.

A

Last but after that, I set up a webhook in captain that captain can can trigger the deployment of argo and last but not least, yes, I labeled. There are our applications with the captain information, as we saw before. So I added the information about stage, project and service in argo city to get some kind of correlation between captain and argos.

A

So yes, as we see the evolution takes a bit.

A

A

Wait some seconds.

A

um Yes, uh what I wanted to show you two was the meanwhile also our test finished, so we used k6 for testing and, yes, we see the output of this tool. Here we simply did some did some requests on the on the front view. I love to put off the potato head.

A

What we see here is was that all of the all of the things were successful and in the meanwhile also our sli evaluation is finished um and yes, we also see that it can be promoted to the next stage. So what you see in the sli screen here was that we had six slis.

A

All of them were okay and yes, so I think our deployment was very successful in this case. So let's get back to the presentation.

A

So um what did you see today so at first captain is able to orchestrate argo deployment and we get the post the best of both worlds by integrating argo and cabin. Furthermore, um the integration is simply done via backbox and notification, so there is not yet another integration service. This is all coming out of the box. With captain argo, we simply have to configure it.

A

Furthermore, yes, the deployment is done using argo, cp, the validation using captain, and, yes, you can do some promotion via pull request. This is currently working progress, so there is also a kid promo already keep promotion service, but this will be improved in um the further in further consequence.

A

So how can you proceed at first um feedback is very welcome, so we have a channel for the github for the github's working group in the captain's leg.

A

Furthermore, the documentation about the blueprint of the config of the argosy integration is stored in the github repo, so just try it out make improvements or raise up issues when you find some some problems with this.

A

Furthermore, also this demonstration is available in the github repository, so everything I configured here in argo city and so on um can be found in this repository.

A

You can try it out open issues and, if you've, if you think that some some configuration could be improved or should be improved, um feel free to do it or raise up issues and last but not least, and as I said, enhancements and improvement improvements of disintegration and this mechanism is very welcome.

A

So I hope you had fun in this webinar and I hope you will have fun with argo, cd and captain and yes thank you for your attention and hope to see you soon. Bye.