Cloud Native Computing Foundation KubeCon + CloudNativeCon North America 2019 (San Diego), 22 Nov 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Intro: CNCF CI - Lucina Stricko, Vulk Coop & Denver Williams, Vulk Coop

Description

Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io

Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects

Intro: CNCF CI - Lucina Stricko, Vulk Coop & Denver Williams, Vulk Coop

The CNCF CI status dashboard -- cncf.ci -- provides a third party validation of builds, deployments and end-to-end testing for CNCF’s Graduated and Incubating projects. The CNCF CI status dashboard continually validates each CNCF project, for any commit on stable and head, running on Kubernetes clusters which are provisioned to a bare metal environment. The results of each testing stage are published to the cncf.ci status dashboard. An Intro session will give an overview of the cncf.ci status dashboard’s key features, goals, technologies used, and allow time for Q&A.

https://sched.co/UajL

A

My name is Lucena strick, oh I'm, with volt, cooperative and I'm here with my colleagues, specifically Denver Williams, who is the co-founder of the CNC FCI initiative.

A

So the session is 35 minutes in the session. I'll introduce you to the folks working on the CNC fci dashboard, an intro to the initiative, including some goals, key features, roadmap and an overview of the technologies used about midway through we'll have plenty of time for questions and answers.

A

So a little bit about volt cooperative. We are a software cooperative in business. Since 2013 we are based out of Austin Texas and have folks working with us from all around the world and in Austin Texas we lead to meetups Austin software cooperatives, Meetup and the open source axes. Axe-Throwing meetup, here's part of the team in Copenhagen, @q Connie, you pictured, we've got Taylor Watson, Denver and myself and not pictured. We've got Ashley, we'll Josh and Robert also on the team.

A

So at a glance this is CNC FCI.

A

The project consists of a CI testing system, a status repository and that user facing dashboard. The CI testing system can use existing build status from a project CI system like Travis, CI or even github actions, and it can generate its own build status and artifacts. We validate the build and deployment status of each projects on kubernetes, stable and head on x86 and arm architectures on packet servers, the status repo collects the results and then the dashboard displays them.

A

Some of the goals include complimenting the CNC F landscape, which helps enterprises get started on their cloud native journey, and we hope that this dashboard shows graduated and incubating projects in action.

A

We also like to promote C and C F, hosted graduated in incubating projects to attract more projects to C and C F, and on the dashboard we demonstrate the use of cloud native technologies on multiple test environment configurations by collaborating with project maintainer x' to add and maintain projects. We hope to contribute to a scalable project ecosystem and the visible dashboard can facilitate conversation, and we welcome feedback from the community.

A

And finally, the dashboard provides a vendor, unbiased validation, of build deploy and end-to-end tests for projects, and it's available as a program management service for scene. C F hosted projects.

A

The dashboard displays graduated and incubating projects, including eight graduated and 15 incubating ideally, and we also show the onap s o project v1, Amsterdam release.

A

Ona v1 was added at open networking summit in Los Angeles and since then, onap has significantly improved their deployment with kubernetes and their latest releases. The CN CF CI team was able to reuse the knowledge that we learned while working on own app for the CNF testbed initiative.

A

So our key features of the CI dashboard support all of the goals. Above first, we aim to be project centric. The projects are at the front and center at the CI dashboard. We show the passing and failed status badges for the stable and head releases. We can reuse, helm, charts for the deployments and the end-to-end tests provided by the projects.

A

Next by collaborating and increasing engagement with the project, maintainer x', we can add projects quicker and reduce the level of effort for maintenance by using upstream build status and artifacts. Maintainer x' can follow the contributing guide, to update their project, to release, name and build information and also help provide home, charts and ends and tests for the test phase. And third, we show cross-platform testing on a neutral, bare-metal testing environment.

A

We provide configurable test environments, currently supporting two kubernetes releases, with room for more and two architecture: types x86 and arm, so we can walk through the history of where we've been and where we're going on the CI dashboard.

A

Here's a review of v1, formally known as cross cloud CI. This was released in early 2018 shortly after Q Khan in Austin Texas and the first iteration included support for several cloud providers. It showed the status of three pipeline stages, the project build provisioning, kubernetes and deploying on to the provision clouds. Some of the lessons we learned with v1 was that we found little difference in how the projects were deployed to the provisioned kubernetes cluster across the various clouds.

A

We also found that different special interest groups, working groups and certifications had been started to work on kubernetes conformance the dashboard had more clouds than projects, so we decided to change direction.

A

V2 was released in February 2000 nineteen with an updated user interface and the primary focus was to put the projects in front and center.

A

We reduced the number of columns from 11 wide to 5. We added a test environment section to support multiple configurations running on the bare metal packet servers.

A

Just last month v3 was released and in it we changed how kubernetes is provisioned. We were using a custom, kubernetes provisioner called cross cloud, and we decided to switch that out with community tools like cube spray and cube ATM. The way the dashboard works. Is it refreshes daily at 3:00 a.m. Eastern time the badges refresh they show running when the jobs are running success when the job is past and failed.

A

If there's an error for the test environment area at the top of the screen, the CI system provisions each kubernetes release using container D as the container runtime onto packet servers.

A

We added arm support, arm, join C and C F this year and was announced as a Platinum Member just this morning. The CI dashboard is a great way to visualize CN CF projects and kubernetes working on arm platforms. Arm support has been added to the kubernetes, stable release and the latest commit build and provisioning stages.

A

Our support was also added to graduated projects. Accordion s envoy, fluency and Prometheus for their build and app deploy, stages clicking on any of the status badges on CN CF CI will bring you to a get lab job that shows the status and any error alerts that may have happened. This one shows Cordy and a stable, build for ARM architecture. So, on the get lab you can see we're on the cordilla Ness project.

A

The architecture is showing as arm 64 and that also shows core D & S version 165 so feel free to check it out. Cn CF CI.

A

Let's take a look behind the scenes at some of the technologies used for the CI dashboard.

A

The CI CD platform under scenes, the up CI, is get lab. The test environment is provisioned with terraform and cube spray. We use helm for app deployments and get-and per project, diamo configurations for the build and release the front end is view j/s and the status repo uses, elixir and Earling.

A

The provisioning stage we provision test environments, packet, x86 and arm servers are provisioned using terraform, the kubernetes, stable and head builds run and then cube spray provisions. The clusters on to the packet servers. Then, of course the dashboard shows the results for the build stage. The see a testing system can integrate with external CI systems like Travis CI, as well as generate its own, build artifacts internally, which is what we're currently doing for the ARM architecture.

A

The sia Testing system creates or retrieves build status and then displays the results under the build column for app deployments on the provision test environments, the CI testing system collects the build, artifacts uses home charts and then visualizes the deployment status on the dashboard and for an in-depth look on how we implemented some of these stages and technologies. We'll have a deep dive tomorrow morning at 10:55.

A

So that's an overview of where we've been and here's what's next on our roadmap.

A

Will collaborate with Vitesse on adding their newly graduated project to C&C, f, CI and we've started the external CI plugin for github actions to retrieve the build status for projects using github actions will also resume efforts to use external, build artifacts to populate the deploy column and then start adding the ploy for linker D to Jaeger or tes and we'd like to support stateful URLs for sharing and tracking test results by day, and in order to do that, we'll need to start saving history we'd like to collaborate with the projects on populating the ends and test column, which is all n/a at this time and will add new projects using the github actions, integration and then start on the third integration.

A

Fourth integration. So if you have any ideas of what CI system you'd like to see next, please let us know we definitely welcome your feedback and collaboration. There are several ways that you can get involved, we're ready to add the graduated and incubating projects on the cross cloud, CI repo. We have repo set up for all the graduated and incubating projects, so please see the contributing guide to get started. We welcome your pull request or issues.

A

If there's any issue or question that you may have, you can also subscribe to the mailing list, join the slack group and attend the CI. A working group calls upcoming events. Tomorrow we have the deep dive with Watson and Denver at 10:55 on how to add new projects going into code review of that and Taylor and Denver will be presenting on the CNF testbed initiative on Thursday afternoon, and the next CI working group will be the fourth Tuesday of January at 12 noon.

A

Pacific time you can stay connected with us on github, twitter, slack email, you name it and I think we have some time for questions with Denver. So please welcome to the stage Jennifer Williams co-founder of C&C FCI.

B

So what since gonna run one around mother mine? So if you have any questions, oh.

C

Hi Demarest weeka, um quick question ice I, don't know if the part of the actual products handled is like the different communities they handle this, but I didn't see any type of static code, analysis or scanning of the code or I. Don't know if that's a part of this project, or is that something that you tell them to do and then you guys worry about the end result. So.

B

With the new integration way of just interested in the end, results they'll configure their integration and then, when they see our jobs around we'll go through at 3 a.m. and check the status of what happened.

D

I have two questions. One I want to. Thank you all for your incredible work. I did. This is more like curiosity than like a technical question, but how do you organize all of this and get people to collaborate on it and get funding for it and pay for your infrastructure and stuff like that, so.

B

It's EF projects, we've got funding from them and we've got a working group. This yeah working group so most of its us reaching out to projects or they reach out to us and then we'll get on a call and then collaborate with them on how to add themselves and help cool.

E

The infrastructure is the soon CF community lab, so that's available for any project to use.

D

Unpack it yeah cool and then kind of building on that I have been involved in elephan and I'll have related projects that did integration testing with other projects in the past and kind of treated projects now I'm working on more kubernetes, centric ones, submariner for networking. It's still like a very new project, but I have, in the back of my mind in the future that you know this is something that we would eventually love to integrate it with and I. Imagine other people have similar questions about how to integrate with you guys.

D

Maybe you could talk to that for a second I saw some contributing links, but yes,.

B

We've got some contributing links there and they cover, because with the new integration, it's mostly self service. In a way, there are still some steps that we need.

D

B

If you follow the contributing guide, you can pretty much get most of the way through it and then ping us and then or if you get stuck or.

F

Come to the deep dive tomorrow.

E

So one of the things that we'd like to do and we've been talking internal, is making it easier to have the dashboard use for more projects.

E

Own apps specifically has a ton of components and we're just showing a couple of them on here and really what they need is like a dedicated dashboard, and you could look at that for many projects and we're planning on having one for the CNC FCI infra, like all of the components that are running this and actually be showing that so that people could see the status of those Bilt's and as part of that effort, which we're planning I think it'll make it easier.

E

If there's a project that you say, we'd like this, but it may not fit with say the scenes, if graduated and incubating that there could be another dashboard and ends up using all the same software, and we definitely want to make that more usable for everyone. I think gitlab they're wanting to make like the base platform more usable and they're interested in talking more so.

F

Hi, this was really amazing, like I know, there's a lot that goes on in the background, the I'm very interested in the and aesthetics repository, sir, because it seems to sort of take in the different projects track. What version it is based across I, don't know what the word is.

F

What do you call something where you have a cross-section of things and match them? Yeah yeah, the matrix of the different projects? I, don't know if this is available for reuse elsewhere, because it sounds like something that would be very useful in a micro service architecture where you have different things being deployed across different staging environments, for example, and being able to see what the status is.

F

What versions are in different places and so on? Yeah.

B

That's it's definitely something we always kept in mind when developing it, because we always knew would end up doing matrix, testing testing the different versions, so the status repository is open source and then it has a little logic for tagging or looking at the metadata on the pipelines and going okay. So it's this job here and it's linked to this other pipeline and cross-linking pipelines and that's how it can on the front end map together what version is being tested and where it sits within the matrix?

B

F

Can I find that information.

B

On our github cross cloud, CI is a repository, that's public for the status repository hmm and I'm, pretty sure if you just go to the top-level repo this links to that as well. Okay,.

F

B

All right, I think that's fast. Thanks for coming.