Kubernetes Meet Our Contributors, 7 Mar 2018

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Meet our Contributos: Ask Us Anything 20180307

Description

Check out information on this program and other mentorship programs here: http://git.k8s.io/community/mentoring/meet-our-contributors.md

A

All right, hi, everyone welcome to the second edition of meet our contributors extra special day. Today we actually have two slots today: four different time zones right now, it's 7:30 in the morning on the west coast, bright and early for everyone. We have many awesome contributors on the line right now and we'll get introductions started George. Why don't you go first? George's? Actually, are you streaming guru and he's going to be on the ones and twos for us today? George, let me introduce yourself hi.

B

I'm George caster I'm a community manager and help deal if you have any help getting started, contributing I'm working on the contributor guide. I'll, mostly be monitoring, slack and streaming today and I hope to answer all of your contributor questions and.

C

D

My name is dimps, my name is Radha long I work for Hawaii and I mostly work on things like cloud providers, and you know some testing and things like that.

E

F

Hi, my name is Aaron Berger. I work for Samsung SDS I am a member of the steering committee and co-founder stake. Testing and I do a lot of things related to contributor experience. If you see my emails about turning stuff off for any staff.

A

And Jennifer.

G

Hi I'm Jennifer Rondo I'm, a technical writer at hefty, oh and I, contribute to the upstream committees, documentation, I'm, a maintainer there I've been working on release, notes and upstream documentation since version 1.8. You know when my third release I'm, not sure what 1.11 is gonna hold for me, but we're doing some pretty cool things with trying to get sort of doc standards in place and moving forward. Making that contributor experience better for people, whoo and Ilya I.

H

Am the yet from the works based in London, my github entry to Nick, nameless error, developer I've been involved as kubernetes to to varying extend throughout the time and that can started in 2014 I. Think it's been a while and yeah for in various projects, for example, cube a DM Pennines across the lifecycle. Also I know Blake's about networking and there's other things: vodka, bridges, yeah.

A

Very cool. Thank you, everyone again for joining today. My name is Paris that work at Google, I'm, a community manager and I will be your host answering and asking questions for those that are joining us. If you'd like to ask a question, we will be collecting those in the meet our contributor slack channel on the Cates instance and also on Twitter. If we, if we can get any Twitter questions and hashtag, there is k8s MOC abbreviation for kubernetes and made our contributors, awesome abbreviations all right.

A

So let's do our first question, which was queued up yesterday, and this is sort of a user. Slash, contributor question. So that's why I think we should just go ahead with it. It kind of flows into both testing categories. So this individual is interested in hearing about your development workflows.

A

They are fairly new to the project and, having a hard time, reconciling the contributor Docs, they would like to run a single unit test as fast as possible, but building be a doctor based on the bill took a very long time. They are interested in what the fast feedback development approach looks like.

F

So I can answer as best as I can so I guess, posting links in zoom chat doesn't really help, so maybe I can post this in slack later or something, but essentially I'm. Gonna share my screen and I guess as a quicker way of doing this. So over here I'm in the community, repo contributors, devel file is called testing and if I go down to run specific unit tests in a package, this gives me the command to run like just the tests validate pod unit tests.

F

If I don't care about running just a specific unit test, I could use the what to say just test the specific package so on and so forth. You I believe if you run make tests, it doesn't actually invoke the docker build. It may be incorrect on that, but one of our testing conventions is that all unit tests should run and pass across all the three major operating systems- Linux, Mac and Windows.

F

So if you're running into unit tests that don't pass for whatever reason we have above it, we should fix that we're working on funding ways to address that within CI. But you shouldn't technically have to run the build in docker in order to do unit tests, so there's that next, the happy path isn't the greatest for your initial build, because you kind of have to build everything. But then incrementally this may test. Things should only build just the test files that you need to when you're running and then going a step further.

F

If you're interested in wandering into a slightly more complicated build system instead of using make, there is possibility of using basil, which is significantly more intelligent about exactly what needs to be built in. But that's something I personally am not familiar with and would like to see a little bit more polish applied before I would recommend that to new contributors.

F

That answer the question. Yeah.

D

I, don't I can add a little bit more color, so even if things are running using so, let's start from the beginning. So the reason why go test directly doesn't work is because we generate some files and these files, for example, if you just do a make for the cube, CTL or you know, cube command or anything like that, so the first time it runs, the files are generated and laid down on disk. So the second time when you run you know those files are not regenerated. Unless you wipe them out.

D

So suppose things are very slow, then what you can do is you can run make tests with what that Aaron showed. First, you know just prime the pump basically to make sure that all the files are generated on disk and then you can switch over to go test V and that works most of the time as well.

A

Alright, thanks to you, those for that looks like they have another testing type of question, they're very interested in hearing how other developers are performing manual. Integration test example: they've made the change, built and tested the changes and now want to deploy in an environment. They can run their custom fire in areas in I'm interested in hearing what your specific environments look like and any tooling that you may use and any caches.

E

Tips, etc, that you have for this process.

F

So again, I'll start because I suspect games might have a better answer, so I want to give the crappier one first like this is not something I've had to do in a while, but when I was doing, development I was in a place where I had access to a cloud provider and I needed to make sure, first and foremost that my changes worked within that cloud provider. So testing locally on my machine was not sufficient for me. So in my case I work for Samsung.

F

We have a tool called Kraken and I used that to stand up a cluster with a custom build. My impression is that you can do the same thing with the cluster scripts and kubernetes. If you cube up or whatever it'll use, whatever happens to be present, either a downloaded binary or something that's been built. If I put on my stake testing that I can say if that okay, another thing I'll share my screen about real quickly.

F

So there's a page in the community repo called e to e tests, alright, so wow that window doesn't need to be there. Okay, so contributors develop e to be tests and and testing and kubernetes we're going to go down a little bit here, and this thing here talks about how you can go run this e to e script, with a variety of flags such as build tests, whatever whatever so. Basically, this would create a brand new cluster and it wouldn't do anything else and now I've got a coop CTL.

F

That's pointed at that cluster and I can do things for that cluster. If I so desire, and then I could repeatedly run tests with specific arguments to test just specific things, and then you know over in another window, whatever I can go actually poke.

B

F

The cluster manually, it's been a while, since I've used this route of doing things, but I believe you can also use like a push or an update thing to like, with the notes, still up push out a new build, but that's I may be making stuff up or I may be telling you something. That's you're too old. At this point, and then just one last plug to make is this e to ego thing is actually a shim around program called Q test.

F

So you can rest assured that if you're using that Ichigo thing you're using pretty much the exact same thing that we used to run a to be tests in kubernetes CI- and it's got, you know very, very similar flags up to stand-up cluster test to execute the cluster. And then it's got a couple of things to say: you know, use a build, a fresh copy of kubernetes and then the color cluster.

F

With that- or maybe you know once I built it I can stage it to a bucket so that other things can download it from that bucket later, which staging to a bucket and then downloading. It is usually the approach I've seen a lot cluster deployments, we stink, um that's all I'll, add damn.

G

It anyway in on that one yeah.

D

So let me share my screen.

D

So uh can everyone see this yeah okay, so this is the local up cluster shell script. All you need to do is do a git clone of the cuban artists repository and then just run hack, local up, cluster dot, Sh, that's it and what this does is it builds everything that it needs and it it actually starts the various processes that are associated with cuban artists.

D

It also prompts you to you know, download it CD, if, if you don't already have it and and so on, so essentially the why this is very important is what are the other alternatives? The other alternatives are to actually use cube, ADM to stand up cluster or use any of the other tools that Aaron was talking about, but this one you can just start Ubuntu VM git, clone communities, repository and run hack, locally up, cluster dot, SH, and that's it you're done, and you can run it as a daemon.

D

There are many flags within this script, so you can go look at how things are you know connected, so essentially what we use this for is, if you need a single VM deployment. That is easy to start easy to tweak and quickly. You can make changes and deploy.

D

This is what we typically use at least I, typically use and I highly recommend, taking a look at look at this one so based so now that we have a local up cluster, then how can we run the end to end tests, for example, that Adam was talking about so the main thing I wanted to show here was over the last couple of weeks we added support for a new set of jobs in sick testing to use this this local end-to-end test. So essentially you go to the KS, desperate app spot.

D

You can see the all the tests that we have, and one of them is the local end to end test and this one. Actually, if you, if I click on one of them, you can see that there are two failures, and this is because we are still bootstrapping the local up cluster. It's not fully compliant I'm, trying to run the conformance test here and two of them are failing and we have PRS in progress for both of them.

D

So when that's done and when one ten ships will essentially have a you know, testbed local testbed, so how does this actually work? This works by local I wanted to show you that all it does is it uses local, uh clustered or SH sees.

D

Me I'll try to bump up the font, so you can. So all it does is: does local up cluster and walking through what it does you can take a look at?

D

It builds all the generated files that I was talking to you about before, and then it builds all the all the binaries that are required right and then it pulls out a few more ABD stuff from Ubuntu and then it essentially starts API server and you can see the logs are placed in a temporary directory and then it tells you how to use cube config and how to set up your.

D

You know cubelet configurations and then essentially I'm running this as a demon. So there is a flag called demon equals true or false, and essentially, if you don't want to run it as a demon, run it as in a separate terminal and then you can use cube CTL command from one of the terminal. So what does it do so once it starts the local end-to-end tests? All it does is it runs.

D

It runs the end-to-end tests if I scroll down you can see these are. This is typical output that you see from the end-to-end test and if you want to see the failure, so that's the failure and you can see that you can see the logs that are associated with that test just before the failure. This thing now, how did? How did I actually put this together right, so I can show you what I did what this internally does right. This CI job essen is all it does.

D

Is it runs a specific image that was pre built by our SiC testing. Folks called the tucán latest master and you can assimilate what we do in that CA job just by doing a docker run and start a container and run the test. So if you look at the so that's the docker command for that I have a gist.

D

If you go to Gstaad github calm and do cube that cubed sh, you can see that guest just so, oh and in inside the shell script run test dot sh all I do is like start docker and then do a git clone, and this is exactly what Aaron was showing you earlier, which is a cube test command and the most important thing here is the deployment and the provider.

D

If you set it to local, then it uses the local up cluster to start using up down means, bring it down when you're done and test means you run the specified test, and here using this we are essentially saying run all the tests in a single thread. Don't run it in parallel and we are running the conformance test here and instead of conformance you can run. You can type in the exact specific test that you want to run in the heat end-to-end test. So that's all it does so you run the docker container.

D

The docker container starts off and then do you do a git clone and then you do a Q test. Cube test starts the local up cluster and then runs a test and reports the values. So it's pretty simple to you know, get started and kick the tires. I hope that helps the one other thing that people ask me is like how how do I figure out? What tests are there and how do I run those tests?

D

For example, they were audit tests that we're failing how how do I figure out which tests we're at where the tests are right. The tests are in our Cuban, it is Cuban, it is repository if you go to if you go to tests, end-to-end tests and most of the tests are below here and let's just pick something in storage, for example. Right, if you say, if you go here,.

D

Let me search for conform and it right conformance. It is the method that we use to annotate to make essentially make the call to set up things. So, for example, this one is conformance test and if you copy this string and stick it in where I had conformance, then it essentially runs this one specific test. That's another tip that is usually helpful because it takes a long time to run all the tests. Sorry layer, you were saying something: okay, so that's what I wanted to show?

D

Maybe one or another thing that I wanted to show was okay, say: let's say this one right now: how do I easily search for things in a Cuban a test repository, so this is one utility I. Have it on my personal I run it on my personal hardware, so you can use it too. So, for example, so it's easy to search for anything you want to across all the Cuban artists repository.

D

So this is very useful for me when somebody puts up a random track, stray stacktrace or a message: that's coming out from communities have no idea where it is then I just plug it in here and then go looking for stuff. So that's what I wanted to show you? It's really for debugging as well.

A

Tempest was extra helpful. You definitely went out of your way to answer that question. So thanks so much appreciate it.

D

A

Right, um we actually have two testing questions in the queue as well, but I'm going to skip to with some other testing. To give you both a break we'll go to somewhat of a general ish question Danny and the Chad wants to know if there's any SIG's looking for new contributors or any new participants.

A

So this is a good time if you are in a Sager working group that needs new folks to help out too, to talk about that, but then also in a more general sense, you know how do new contributors find find out if other SIG's are looking for folks?

A

Anyone you want to take that one Jennifer I know maybe docks docks.

G

Is always looking for contributors? We welcome new people, it's the start of our weekly meeting and when no new people show up we're sad.

G

It's it's also true that we get a fair amount of rotating in and out of, cig docks. There's a core group of maintainer, that's pretty stable and then folks who drop by when they can, and so we appreciate involvement in contribution.

G

Whenever and however you can, you can manage to fit it in and there's all kinds of work available to do on the docks, there's, definitely technical stuff that always needs to be written and reviewed, and there's often you know, usability stuff if you're new to kubernetes, you know it's kind of a classic point. You've got fresh eyes on things that don't work.

G

There's actually been a great conversation going on in the slack channel and about installing cube petal and cube ATM and ways in which say the docs don't quite meet people's expectations and I've copy pasted. The whole thing out, and as soon as somebody gets, a PR in I will go, take a look and see whether navigation or content reorg can help. So it's a really nice.

G

What's going on in the channel right now, it was a really nice reflection of the ways that you can contribute to Doc's whether you want to join the sig or you know, sort of do a drive-by. It's. It's really really excellent. Awesome.

A

Oh yeah, do you know in many groups or look yes.

I

uh Denise across the lifecycle he's always looking for new contributors as well.

A

And they connect with with that group and and look.

I

So a good place to start would be the Sigma plus the.

H

Lifecycle, slack Channel or the weekly meeting, we, we have two main weekly meetings, one on Tuesday, which is morning like to think it's ten o'clock, Pacific time on six five o'clock, Greenwich Time and there's another meeting. That is that is happening today right after this meeting, which is a QA DM office hours, which is, which is probably a good one too, for new contributors to join all.

C

D

H

Be found in the community, we both miss across the lifecycle page week immediately good.

A

Plug yeah I also know that Service Catalog is looking for new participants and I actually just found that out today on Twitter, so connect with Service Catalog folks on slack or their mailing list stores. Illya just mentioned. Their information is also within the community repo understate list, as are all 35, plus special working groups or special interest groups and working birds. Also contributor experience is also looking for help with creating the best possible experience for contributors. We actually have a meeting today in an hour and a half ish if you'd like to join same deal.

A

Our information is on the sick list within the within the community repo anyone else have any other plugs for groups that may be looking for new contributors, I.

D

Can just talk in general a little bit, so it depends on your interest right. The people who are there is a lot of things. You can self discover. The main thing that you look is in the community repository and in the Kuban. It is repository you can, you can go, hunt for information there and the other thing that I would say is: go look at the snack slack channels and hang out and you just can be a fly on the wall. You don't even have to say stuff.

D

Just till you get to know the people and things like that so and I usually go to Sigma Tings of different kinds and for like two three weeks: I, don't even say anything maybe more. But then I can you kind of like absorb how they work and what they do and see where you are comfortable and with. Where do you think you can help? It doesn't have to be a big deal, even if there are small bugs that you can triage and small PRS that you can review.

D

That's the best way to start, and some of us don't really know didn't really know, go when we started looking into to monitors. So we all learn on the fly. So you know do the start somewhere and you know, keep going that's what I would say.

F

Just gotta put in a plug for sake, testing, I guess but I think everything Tim said is great, like I personally, sometimes I feel it's almost like the bit of any inside pattern. That folks feel they have to first join a cig and then figure out what to do. Although I understand that it's really useful to know that there are humans that you can contact, and so I do think that lurking it's like in slack channels to get a feel for the vibe of this sake would help so my plug for sake testing is.

F

We have added I, believe the most custom slack emojis. So we we are super heavy slack participants. It's a great place to hang out and I think like little tiny contributions that we would really appreciate. We're not doing a super awesome job at applying the Help Wanted label to things that are pertinent and relevant, but.

B

F

Getting there and I believe we are looking to add four new contributors label back in this is a discussion. That's ongoing with contributor experience, but like one of the tiny things, for example, is we have a lot of end-to-end tests that fail with expected error timed out waiting for condition not to ever occur? I'm like what condition what was the timeout? Where did this happen? Why did this happen and like stuff as simple as adding it slightly different log message would help tremendously with this anti-pattern there's a few small things.

F

There are lots of them to make there's a Help Wanted issue for it in the kubernetes repo, and we definitely welcome the help. Awesome.

A

And George I think you got off of our god on video. Rather, did you want to plug something.

B

Yeah I just want to say it's totally fine to not pick a cig and then just shop shop around just attend a few six. Sorry I got construction everywhere, go to a few cigs, a hey, even if all you do is take URLs from what people are talking about and paste them in chat kind of see. If it's for you and then just just want, walk the earther and then figure out what you want to do. That's fine, too yeah and.

A

Also shout out to George Gwyn, Tim, pepper and others who are actually working on a contributor guide as well, which you can find in the community repo under contributors. There is extensive work with that and making it discoverable and also a future developers guide as well. That's more comprehensive!

A

So if you know if anybody on the line wants to start contributing honestly I think that's a great place to start. You poke holes at our process and also figure out what developers need as far as documentation is concerned and again that work is coming out of contributor experience. So, if you're looking to dig in there join us next question and then what we're gonna do is something special with Jennifer. Jennifer is actually going to take us through a doc review. Jennifer, you still game for that.

A

All right, cool one, more question and then we'll get into that.

A

So Ilia do contributors generally build with docker or without oh.

A

We can't hear you or can anybody hear oh yeah yeah.

H

All right, yeah I, should be able to hear now what I'm saying well most of developers who who use a Mac, laptop or women's lots of perhaps will probably end up using docker right, so I think if you're developing on Linux you, you may get away without it, but sometimes it's just easier to do not have to configure your machine to have all the dependencies.

G

Anybody else wanna weigh in on that question.

G

F

Right yeah I mean like I, was alluding to earlier. If you're just doing like unit tests or integration tests, you can totally get away with developing and building locally. It's if you need to. Basically anything that involves kulit and actually testing it. Live requires Linux, right, there's, there's not a couplet that natively supports orchestrating containers on Mac we're getting really close to having a couplet that can do that on Windows, which will be amazing but yeah I. Think.

F

Ultimately, if you get to the land of having two fully-fledged integration testing with the whole cluster, you know: you're gonna have either Linux as your development, environment or virtual machine someplace, where you can run all the things.

A

All the things all right so now this is the second segment meet our contributors. We're. What we're trying to do is either a live peer code review or a live peer, Docs review, and since we have Jennifer Moines today, she's actually going to take us through a Doc's review. Jennifer.

A

We do have more questions, so do not feel like you need to to squeeze everything or even elongate everything into 30 minutes. So when you're done you're done- and we can ask these folks some more questions, cool.

G

It's 30 minutes worth of doc. Reviews would not be sufficiently interesting to anybody. Don't worry, I'm not going to inflict this on you. What you have is I think like three different for requests, queued to are already merged and one is a waiting review just to show you the range of things that we deal with, because people come to Doc's with all kinds of assumptions about how to contribute what's valuable and we we get all kinds of things.

G

But if you don't know you don't know so when we fiddle just for a minute with sharing screen because I don't always get this set up right.

G

Okay, have I got this, so you can see I really yeah, okay, so I did check in with the folks who submitted these. These pull requests not that they're, not a matter of public record.

G

They know that you know we're talking about this, and so this is a new doc or you know, contributions to an existing doc but fairly substantial new content for the upcoming release. So it's additional technical information and one of the things that the cig Doc's maintain errs do not have is expertise in all of the parts of kubernetes. We depend on outside technical reviewers to verify the technical accuracy of a lot of PRS for minor things or for things where we happen to have expertise, because we're involved in other SIG's or working on related projects.

G

We will do tech review ourselves, but our standard process. You know where we've got substantial content additions just to make sure we have both a tech review, LD, TM and a doc review, LD TM and in this case on what you see is all technical changes, there's actually some tech there's there's a technical conversation going on between two contributors to this feature. Actually, that's the folks who added this there's side conversations that I happen to know about, because I'm running the docs for on the one-to-ten release, and so you know, there's some night.

G

There's some interesting bits here. Right I mean the comment. I really like here is inconsistency between implementation and design document right it happens and that's one of the reasons we need tech reviews so that people who know where the different moving parts are actually happening can check up on each other. On this review has so Joe heck I missed one of the sig ducts maintainer and he's simply given a clean bill of health from a Docs spective.

G

We sometimes get pickier and so with regard to docks, ldt ends, and so let's see this, this one is still open. There's a little bit of sort of wording.

G

Wording stuff going on here in tech review, but what I need to go in and do now is to provide some feedback around conformance to the docks style guide, and we are experimenting with how much we ask for compliance with the style guides tip. Historically, we haven't been particularly insistent because we simply have had to larger volume of pull requests and we need to get the content in and we cannot worry about whether people use Latin isms. That's this via right here.

G

The annotations makes the technical writer in me twitch, but if people are going to understand what it means, we're not going to worry about it too much depends on where we are in the cycle and what the PR backlog looks like honestly, on how much feedback you're going to get about compliance with the style guide, and we are in the process of reviewing how much of the style guide makes sense to adhere to going forward.

G

The style guide exists for a couple of different reasons. One is to make sure that the English is as accessible to a range of native languages as possible, and the other is to provide the kind of I guess, authoritative, reassurance that consistency can provide so that things get named the same way in every. In every case. We don't have that in all kinds of ways across the doc set one of the most obvious, perhaps which I don't think is breaking user experience, but you know, makes the copy editor in meat, which is capitalizing.

G

The names of kubernetes objects right. Some people do some people, don't some reviewers catch, some don't and whether it matters is really kind of a matter of time right. What? What do we need to get into the docks as quickly as possible? So, as I said, you know this particular this particular pull request. I've decided I'm going to see what happens if I get a little pickier with regard to compliance to the style guide, but I'm also going to be very careful about how I make those comments not being hard-assed or insistent about it.

G

Some of this is education right. Some of it is getting. People used to noticing things about the way they write, and some of it is sticking out that we just don't worry about certain kinds of things that other style guides might worry about. We try to keep it light weight in the first place. The kubernetes docks style guide is the smallest style guide, I know of for any docks project and that's to make compliance easier, but if it's not easy enough, we'll make it later right.

G

On the other hand, if some things, if we need to insist on compliance because there seem to be problems, we will try to adhere to those guidelines. So that's actually a place where new contributors to docks can provide feedback. Also, most of the time people don't look at the style guide, we're trying to figure out how to get people to pay attention to.

G

But you know it is a thing that the maintainer --zz try to pay attention to try to figure out the best way to sort of provide both guidance and appropriate levels of compliance in this I. Think it's not all that our reviews aren't that different from code review, but there's obviously some important differences as well, and then I'll show you one more thing um like I said: Paris I was not going to take a whole lot of time. Here.

G

We've got oh I had a few more, but here is here is a pull request on a completely different order. This one is huge.

G

So these these first bit are just little bits and then we've got a whole think this is a new topic so dealing with high availability and I'm, showing you just the initial, the initial PR and one set of comments. But in fact, there's extensive review. We've got three technical reviewers on this topic already and when I go in to provide a doc review, it's going to be significantly lighter weight than it would be on a smaller PR. This one's been hanging in there for a while.

G

We need to get the content into the docs if I decide or one of the other maintainer x' decides that. Well, really it could use some better conformance to the style guide for usability purposes, we'll go in and make those pull requests ourselves.

G

We don't have to insist that absolutely every contributor to the technical side of the documentation, you know sort of be as nitpicky as we might be, but again it's a matter of available resources and we're always having to sort of slice and dice here at a certain level, anybody can contribute to the docs anybody can write, but we're also trying to get to that point where we provide the guidelines toward making it more usable and again the conversation about certain how which content you highlight and how navigation and presentation might help that that's in the major contributors, I'm slack right now.

G

That's a good example of where the technical information and content organization or information, architecture and ways of wording things can come together so that people help make the most usable possible docs I'll, stop there. If anybody's gonna know how we review.

G

This is also very much a work in progress, so developing doc review standards, one of the biggest things we're trying to do right now is get through the PRQ faster.

G

This has been I mean it's a huge problem for everybody right, but we have been trying to embarrassed at how they've backed up and we've just as of last week instituted a rotation where the core maintainer is are signing up for a week of being in charge of the PRQ and making sure that we get at least a first response within well right now we say a week, but the idea is with assigning in sort of queue.

G

Tenders on the idea is to get a first response into new PRS within 24 hours, even if it's nothing more than making sure that it's assigned and that somebody is going to monitor the thing we get a lot of lightweight ones that we can pretty much get through immediately. We get you know, we get spell checks and we get grammar Corrections and those things we can.

G

You know, as long as we commit to reviewing in timely fashion, we can get those ones through the queue pretty quickly, but inevitably things like this high availability, one come up where there's a lot of back and forth with review, and one of the things we have to do is just make sure that that keeps moving forward. We are trying to make a better commitment to that.

A

Awesome Jennifer question for you and if there was one thing you could tell contributors about Docs say: like some: you know a tip that or trend that you see that's going awry with. What would that be, so you I'm sure you might even hit spelling actually or something to do with spelling.

A

What's one thing that you would want to tell contributors about about Docs.

G

Try it I think, there's a lot to be learned if you're interested about what makes good Doc's know I do have. I do have something more specific than that. One of the things that I notice about technical contributors to the docs is a default assumption that this is how I like to experience documentation if you can get outside your own head even for five minutes and think about how somebody else, who maybe doesn't have your background or your experience, might need the information that you're trying to provide.

G

You know just you know, especially if you're a technical contributor, you're providing information. You probably know all too well. This is kind of like everything everybody else has said about Newton you about first-time contributors right. You know you, you have a new perspective.

G

If you're a first-time contributor, you don't know all the things, and so you ask the questions that people have forgotten need to be answered if you're one of those experienced people providing us with content or even if you're not try thinking about what it might look like from the perspective of somebody else, just try stuff and shoes for a minute five minutes. It's not easy to do, and sometimes you don't have much information.

G

One of the things that I find most challenging about working on the kubernetes documentation is figuring out who my audience is, if I'm going to go in and either reorganize content or suggest significant changes to a topic, because I think it's badly organized I want to understand who it is reading this content might need a different organization right and I. Don't always have good information about that. You know our audience is broad.

G

It's I mean in the case that this particular meeting or audience is. You know new contributors, but kubernetes users come in so many different sort of flavors and people exploring the docs come in so many different flavors that getting a handle on. That is a real challenge. It's part of what drove the the user journeys project that we launched earlier this year. The new portal pages for the documentation site, where we're trying to address what people coming from different sorts of backgrounds or people who have different roles might.

B

G

First, when they're learning about kubernetes, but that's an ongoing process that we were finding anytime anybody contributed, the Ducks can can help us understand that better. That's just golden.

A

Thank You Jennifer for that that was awesome, and if anybody listening wants to do a live, pure code review or Doc's review for any of our next sessions, please go ahead and submit one. We would love to review it. That was more of the the process that Jennifer takes, but if, like I said, if anybody has something that they want reviewed on the spot by a peer, this is a great place for it. Alright, so I we have some other questions and then we'll we'll wrap for this session.

A

The one of the questions that I received privately on slack are is for all the contributors and they want to know what is a business case that they can give to their boss to start up to upstream contributions to kubernetes on work time, especially if they're it sounds like they're already using kubernetes right now, yeah.

D

This is near and dear to my heart. The first thing you need to do is find something that is broken, that you need fixed for your own use case and then that'll get you the initial time for the first step, and you do this a few times, then your boss will say: okay you're doing some important work. That is reflected, then that is how you get into the set of thing. Whatever you work, you do at work.

D

You know me unique justification as well right, so try to fix problems that you have, that your co-workers have in your in your deployment or in your use case it could again it could be something very simple, but that's how you get started. In my opinion, you.

C

Know something add to that.

F

I mean can't give I think again, I think the answer was better for the more concrete direct tactile thing. I say this is somebody who's thought later at the open source, Leadership Summit right now, where that question is coming up over and over in sort of the grander sense like. Why should a company even commit to open source period or why should they eat? Why should they contribute back upstream instead of just using it and I?

F

Think one of the answers that comes up over and over is that open source can be one of the best funnels to your company, both in terms of gaining reputation that your company actually knows what it's doing with the particular projects it's contributing to oftentimes. It can also make the community around the open-source project, one of the biggest fan bases for your company, depending on how actively involved you are in it.

F

So I can say, like the top of at the top of my mind, are the companies that are allowing people to contribute most effectively to kubernetes I. Think it's also encouraging to see such a wide array of companies contributing to kubernetes. So as to ensure that there's no one particular vendor who locks it up and steers it in just one direction and another sort of slightly good thing is I. Think experience with open source projects really opens your eyes to what it takes to get up and running relatively quickly.

F

And you can then take that experience back to your own company and helps with making introductions to companies projects a lot faster. I've heard stories of people who, like once they hire an open-source maintainer who's used to. You, know a readme and really simple steps to build test and contribute to a project, and then they look at the company's internal processes and head explodes, and they may help with that.

F

And then the company can sort of see like oh look, we're capable of onboarding new people and like getting to production significantly faster, because we've adopted some of the best practices for a moment source. So you know that's the longer strategic upsell but I. Think more tactical tactical, like you're gonna, get experience and you'll to find people who will help you out and build a great Network for getting questions answered more quickly.

A

And then George I, like.

B

To tell people consider everyone here as part of your IT organization, like you know, just like you would, if, internally you were writing something like kubernetes internally, except you're, saving a ton of money and it's you know avoiding a ton of mistakes, so I I kind of approach it that way. Right, like you, know, we're gonna solve problems, and somebody else is gonna solve the problem and we're going to get a benefit from that.

B

Even though this company might be in a totally different field, it might be a bank and we're a grocery store kind of thing. So I think there's a case that you can mate or a manager. It's like hey, a bunch of us together is way smarter than you know, as as smart as the people who work here are working together as an industry across horizontally.

B

You know different fields, and things like that ultimately shows I mean because that's how kubernetes came to exist in the first place, so I think that that kind of shows that there's value there and I think the savvy companies recognize that, and you know that ties in to the people want to work at those kind of companies that get it right, because there's there's nothing more frustrating that working in a company where you can't grab someone else. Who's fix the same problem right.

B

If I'm in a silo fixing a problem that Aaron fixed I'm, not happy. You know, whereas, as you know, something like commodity plumbing like this there's there's a good benefit of my employee have or my employers having or our employers having good relationships so that if Aaron fixes something we all.

D

B

We can get on to actually fixing real problems or companies instead of getting stuck in infrastructure gunk I. Guess, that's all! That's all I kind of pitch it to people.

A

Good pitches, you know very good last question and this came through private slack message and it's most likely for Ilya, where European contributors hang out.

A

Meetup groups or other things or events or slack channels, or anything like that that you can share with European yes good.

H

Point um so we kind of don't have a dedicated slug channel I mean we should have one actually, because what I, what I, actually find quite hard, is to sort of find somebody in time zone whoo-hoo, you know be able to just point me in some direction when I, when I need some help and I often end up having to wait until the next day or late and evening to just get some basic questions answered or I.

H

Just ask the few friends of mine who I know are in this time zone so, but that usually means yeah and basically going to one or two people with my questions, so I think I think it probably would be good idea to create slug Channel, but we we do have meetups. Obviously, in most of the major cities like London coverage, meetup was pretty big, but I had to come to zero.

H

Actually, don't know too many contributors who go to the meetup meet up in London is mostly for users right most most of the entities I uses some are contributors, but maybe they pay the camera. They did I mean the contributors who come to the meetup.

H

That I know are not working on comparing this project directly, so so yeah I, don't I, don't really know about other cities, but it sounds like which we should have some some kind of slack channel, but then, on the other end, perhaps you know that there should be some kind of regional flood channels or for each of the major regions which which have significant turns on differences with other ones right, um one more thing, I did to add.

H

We do have the European office hours every every month, which are mostly for users, but contributors are more than welcome to come on and also how about those questions.

A

And the slack idea is a good one. Jordan I are both slack admins and we can make that happen today. We have always been totally okay with regional slack channels. It's just not too many requests. We do have like a China China users group and things like I, don't slack, but if anybody's listening and ever need slack support, slack admins is the channel to go to to pitch any ideas for things along those lines and Ilya will I'll get in touch with you today and create something don't.

D

Forget your pitch cue Connie. You.

C

Yeah I know you as well definitely.

B

C

H

Go to our live office hours.

D

A

You know, and we didn't even pitch to weed there is a contributor summit by the way in Copenhagen for those who are interested in a day before the festivities start, which is May 1st. This is free to all contributors and no, you do not have to have a ticket to Q con to get in, so that's even better for those who cannot attend q Khan all right. Well that wraps this edition of meet our contributors. Thank you.

A

So much to everyone that joined today lots of testing questions lots of introductory questions, Jennifer thanks so much again for the doc review. I will see everybody on this channel. Well, not you not you folks, but I will see the any Watchers again in a few hours here for our next session with new contributors. So thanks so much have a great day. Everybody thanks about.

F

All happy Wednesday things.

A

F

A

Is everyone, but that.