Continuous Delivery Foundation cdCon 2021, 6 Jul 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Lightning Talk: Events of Screwdriver - Alan Dong & Jithin Emmanuel, Verizon Media

Description

Lightning Talk: Events of Screwdriver - Alan Dong & Jithin Emmanuel, Verizon Media

Screwdriver is an open-source build platform designed for Continuous Delivery. It shares the same goals as other CICD tools to increase developer productivity and user experience. In this lightning talk, we will cover terms like Event, Build, Trigger, and showcase how Screwdriver uses Events to make all Pipelines and Builds connected, and use these nuts & bolts to build our UI visualization.

For more Continuous Delivery Foundation content, check out our blog: https://cd.foundation/blog/

A

Hello, everybody. This is alan from screwdriver at verizon media. Thank you for tuning today today we're going to talk about events, the very essential component that connects all the nuts and bolts together in screwdriver.

A

Before I jump in to give you the definitions I just want to talk about, the fundamentals of csd for a moment means continuous integration and continuous delivery. All that start with code comments doing some sanity check test every auto deployment. You can see this circle live for software development right here.

A

Screwdriver supports various of scm, such as github and gitlab.

A

Okay. So now, let's take a look at the nuts and bolts of screwdriver. We have jobs, workflows, pipelines, builds and events, and what are they? A job consists of executing multiple, sequential steps inside a specified container workflow is the order that jobs will execute after a successful build of the first job.

A

A pipeline represents a collection of jobs that share the same source code. These jobs are executed in order defined by the workflow. A build is an instance of the running job. An event represents a single pipeline execution started due to git, commit tag or release or remote pipeline trigger or human menu trigger.

A

Where more terms, your website at docstar schooldriver.cd, and you can check out the links in the footnotes. I will share some pictures that help me to visualize this grocery and the realization of the abstract. When I started devops and heard about pipelines, it immediately reminds me something I used to play when I was a child pep mania, so I took a few screenshots and put them together. Hopefully that can.

B

A

You to connect the dots in the same way, literally, let's start with a simple case. The goal of this is to deliver water from start to end. We have pipeline on the top stage without water field and we have a same pipeline with water field.

A

We lay down the direction of the pipelines and just wait until the water facade to be turned on. Imagine these waters is our code and water flow through the pipelines is our work goes through csct, workflow and an event is what triggers the transformation, usually the means to achieve that is through api calls. In our case, this is usually done by scm webhooks, very intuitive, right.

A

Okay, now, let's take a look slightly complex flow notice that instead of delivering water from one to another place now it delivers from one place to two places and also take a look on the parallel case. We have two starting points to deliver to two separate endpoints in parallel, without interfering with each other.

A

Here is another advanced use of pipelines. Cherry pipelines second pipeline will only starts if the first pipeline finished, so this will only run once this finished.

A

This visual edition of each use case are supported in screwdriver. Remember a few slides back, we talk about jobs and each pipe is considered as a job, and each job can have many steps. Step is executed one at a time as water flows. In now, let's have a look at the real deal.

A

Take our api pipeline, as example as screwdriver. We do what we preach. We use screwdriver to build screwdriver anyway. Here is a pipeline page on the top part we have the header and that helps you to navigate the ui.

A

Here we have an overview of the kernel pipeline status. Commission message status who committed the code? When did it start and how long did they last in the middle section? This is our workflow graph, which is really similar to the diagram. I showed you earlier.

A

This is a visualization of the execution of the jobs on the right side. We have events. Events are grouped together.

A

If they have the same sharp, we can expand it collapse it you can scroll and collapse it and that helps you to keep the events organized. We also highlight the last successful bills and the chart.

A

Let's take this event as one example, this event is triggered from a git commit and it triggers job main and let's take a look at what the job me has. Xiaomi has one two three four five, six steps.

A

After those steps finished successfully: let's come back here: it will continue to trigger job, publish and similarly job publish has 5 steps. After all, steps are finished, it will trigger beta and product.

A

Now, let's take this event as one example: you'll notice, all the workflow graphs changes when event changes. That's because workflow graph is scoped to each event. For this particular workflow graph. It starts not by triggering from git commit like the one I showed you before. Rather it is triggered from a remote job, meaning by another pipeline's job. For this particular case, it is pipeline, 9 and job name is published after pipeline 9 job name, publish finished successfully will continue to run this pipeline and the remaining is the same.

A

So let's take a look. What this remote job pipeline, 9 publishes.

A

So now we are on this commit and check out the last successful.

A

Let's take a look at the build detail. Pages now has.

B

A

Jobs so after this job finished we'll go back.

A

And again, click on triggers that shows hey, which also will trigger after publish job finished and that's how you navigate across pipelines for now, by the way, the context is preserved across pipelines if they are triggered by the same event and inherently their jobs and steps will have the same context.

A

We have more strategies to share data across pipelines such as metadata and build cache. Metadata is mainly used to share unstructured data such as string, integer or json well. Build cache is used to share files and or artifacts.

A

You can check out screwdriver yourself at cd.schooldriver.cd.

A

I want to highlight these following links as our closing remarks, and I want sincerely thank you for staying with me for the whole time. Screwdriver is an open source project and we cannot do this without your support. Please join the slack channel for questions and help. Now we're going to qa session see you there.

A

Here for q a's, if you guys have any questions, please let us.

B

B

Hello ellen.

A

A

Can you talk about the community one uh like how people, if they're interesting, how do they reach out to us? Oh.

B

Sure so the thanks for sharing the git links. So we do have a public community meetup, which happens on thursdays every two weeks. So it's a simple google meet and this uh the links for the meet you can find it in the github page and also find a google doc or you can find our meeting notes.

B

Every two weeks. Thursday is 6 pm pacific time.

A

That's pretty much.

A

A

B

A

B

All right thanks.

B

B