Cloud Native Computing Foundation Contributor Strategy Special Interest Group, 27 Apr 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CNCF SIG Contributor Strategy Governance WG 2021-04-27

Description

CNCF SIG Contributor Strategy Governance WG 2021-04-27

A

A

Ooh, let me change the lighting here.

A

Oh wait: that's the wrong camera! That's why I got it.

A

B

There we go you're looking at.

A

The laptop the laptop camera is my one fail on buying this particular laptop for linux. It wasn't clear that it was meant to be controlled by a proprietary windows driver.

A

A

A

So, okay, um I don't really have a lot today because I think all of us, both of us, have been busy with other things, some of them governance related, but not things. We discussed in this meeting.

B

A

So is there any reason not to go ahead and merge the charter guidance into drafts and send moana and.

B

um Only because, right now it's um it's failing, the netlify check really.

A

B

A

Right and we haven't been able to figure it out because carolyn's been on vacation.

B

Yeah, so as soon as carolyn comes back and tells me, I, I suspect, there's some header stuff, I need to add to this file to make netfly happy. I just don't know what it is or specifically why it's failing and how to fix it. So I suspect it's probably it's probably not a big deal. I just need carolyn's help to do it.

B

Because I could look at the others and just sort of guess at what's wrong, but I think it's probably better just to wait so that we don't do something that breaks the stuff that she's working on.

A

A

And that's kind of it, my you know what my governance time for the past two weeks has been taken up with so likewise the um I mean. Actually, it occurred to me that you know uh messing around with alecto that there's a second way that the cncf could effectively do multi-tenant.

A

That would probably be more manageable for them, yeah than um spinning up individual instances for each project. That's going to use it. Of course. That might also be more.

A

Administratable for them mostly, I just need to finish getting it packaged up and transferred um just you know. We we ran into a hitch with transferring ownership of the electo-ip to the cncf, namely that the dot io tld is blocking domain transfers.

B

Oh, I heard that from someone else, someone else was saying that.

A

Yeah, the dot iotld has always been kind of questionably managed it's one of those things where the uk kind of looted that domain from the island that owned it and um it's in the hands of someone. We don't actually know who now some shady contractor.

B

A

I've stopped registering new dot, io domains and cncf registered electo.dev, but then we need to transfer all the certificates and everything over and I just haven't gotten around to it.

B

A

Wow so yeah the.

A

um Anyway, um so that's what that's kind of waiting on, because it would be nice to offer that as a resource to projects yeah.

B

For sure so much.

A

Easier than the other things.

B

Yeah yeah, I mean we had little tiny growing pains around. Like the you know,.

A

B

The directory structures named and things like that, but even with the little growing pains, it was still way easier than tracking down all the email addresses and managing that or.

A

Dealing with the bounces- oh my god, the um so um let alone the fact that civs is increasingly being used for spam and and as a result, um it's on a lot of people's block lists. um The um so.

A

Yeah, the um yeah I mean. Obviously um we, the nice thing about doing it for k native, is that we've had an opportunity to actually find some of the additional sort of missing bits.

B

A

um Like um the fact that you couldn't mix initially couldn't mix directories between non-electo and electo elections, which I fixed, um I feel like it should deal with formatting problems a little bit more elegantly, and I've been working on that. You know where, if a candidate's profile isn't formatted correctly, it shouldn't refuse to display the election entirely, which is what it's doing now.

A

um It should just not display that.

B

A

B

A

um You know and a bunch of other sort of little stuff during the internship we focused on getting the core voting behavior working correctly, which was the important thing exactly and now it's a matter of adding exception code for all the various ways that people can break it.

B

And is the entrance still working on it or is the intern gone.

A

uh Occasionally, um he's got a real job now so, okay, um he pushed a fix this week, though, for for one of the things that I identified so it's nice I mean, but it's one of the reasons why we choose this. Is it's a flask app so when we get it transferred over and we people start using it, it's going to be very easy to get fixes. Yeah the.

A

Even people who don't particularly do much python can handle flask.

B

B

Yeah, I didn't realize it's written in python. School.

A

Yeah yeah. Well, that was the the choices we offered the internships was, it could be either python or go um because among our team of three mentors, those were the two languages that all of us knew.

B

A

um Because I mean obviously a javascript thing would have also been okay from a public perspective, but we only had one of us who felt we knew javascript enough to mentor it. So.

B

Yeah, I actually do a lot of python, but it's all it's all more on the like data side so, like I know, I know the bits about python that involve like gathering and manipulating data, and those are the bits that I know.

A

Well, I don't know if you looked at flat: flask is pure mvc, so the functionality components really are completely separate from the presentation components which is is good for back end. People like us, because where I would need help is if I have to fix the css or something yes, which I am just no good at at all.

B

No, I I've broken things in really terrible ways: um editing css files.

A

Oh, oh, my god, the um my I had to update my online store for pottery and I just went through this long thing where I was trying to change. The type faces ended up breaking the store entirely had to revert everything gave up.

B

Yeah, that's uh surprisingly easy to do things like.

A

That um and people try to tell me css is not real program and I'm like yeah, okay, great you fix it.

A

If it's so easy, you fix it.

B

So all right cool! Well, if we don't have yeah.

A

Do we have do we have anything else outstanding.

B

I don't think so. I think that's the charter, that's the only yeah.

A

I mean I have other stuff to do like integrate charter statements into the templates, but I have not been able to get to that and it.

B

A

Premature before we well except we're going to merge charter, at least as a draft.

B

A

um I'm also hoping that carolyn is back this week because um I.

B

Was just gonna ask you when she gets back.

A

I don't know I just pinged her in slack because it would be really nice to get the website live before kubecon.

A

I'm going to be afraid that we actually missed the window for that already, because any cncf staff are going to be poorly available.

B

Yeah well, she looks like she's online last week. I think that had a little palm tree.

A

B

Status so she might be, she might be around this week.

A

A

um The um one thing I'm going to start a discussion on and I'll just ping you in slack when I do it is- um and this is kind of more related to the other working group- um is trying to come up with a short list like 10 or 12, metrics, that all projects should look at, because the problem is the various places that we have guidance on metrics cast sort of a wide, because we have the metrics thing that you wrote up, and I want to get more specific because I'm working with some new projects now and they're human bandwidth constrained, and so I really wanted to sort of narrow it down to hey if you're an early stage project.

A

You know here's this sort of list of metrics that you ought to really care about right. You know.

B

A

Would be like you know, new contributors who don't work, you know new contributors with a subset of new contributors who don't work for your sponsoring company um and that sort of thing to add on to that document, um the.

B

Yeah, because we could we could edit that document to be, we could edit it to be a little more focused if we, if we wanted to.

A

The trick is figuring out what those particular statistics are.

A

um I popped over to the chaos project to see. If they had any advice, they did have a recommended list of metrics. There's 140 metrics on it.

B

Well, that's not a recommended list. That's all of our metrics says the person who works on the chaos project. All the time.

B

Yeah, um this is weird, so I I can't actually find the project health one that I did because it was in contributor growth, but it's.

B

B

Am I in the right place, it's a contributor strategy and it was in contributor.

B

A

B

And there was other stuff that was in there right.

A

Yeah there was: where is it all? I don't.

B

A

A

A

Yes, it's under website.

B

A

uh Works for me, I don't care where it is really.

A

A

It does kind of press the urgency of getting the website online, though, because it made the documents much harder to find.

B

A

Without the website being up once the website is up, it doesn't matter because people are not going to be looking at the repo except us.

B

Yeah, so is it a website.

B

There I found it.

A

Yeah because you've actually got a bunch of these here, like in a bunch of the you've, got sort of a start right, because obviously uh the responsiveness metrics are are critical ones for your self performance yeah I mean.

B

So I picked the ones that I thought most projects would need um and I tried.

A

B

It I tried when I wrote this document, so it's not the chaos laundry list of all the metrics you could possibly ever ever use. I tried to keep it really focused on what I thought projects should care.

A

B

So it's probably I guess it's a little more than 10.

A

B

See, depending on how you, how you count.

A

Them well, some of these things are not metrics like do you have a security release procedure? It's not a metric and presumably.

B

There are no public metrics. I have strong feelings about this. It is a metric. It is something that you measure that you make decisions based on: it's not a metric from the standpoint of um something you can get automatically from github.

A

Yeah the um I mean, actually, you know honestly for a more mature project. One of the things I would be looking at is the frequency of security releases, because if you see that increasing, then it shows that you have a quality control problem.

A

A

So because things like you know, for example, uh you know commits prs and issues or commits prs and issues remaining steady or increasing over time. It doesn't really the problem I have with recommending that is. It doesn't tell you a whole lot, because there are all sorts of reasons for those numbers to go up or down that. Don't tell you anything about that whose information to the project is ambiguous. Right issues could be going up because more people are using the project or they could be going up because you have a quality control problem.

A

Yes, um the um and I'd like to recommend things to people that have a more determinative answer like.

B

But the metrics always require interpretation right I mean this is this. Is the thing with the other? Remaining study are increasing over time if they're increasing over time- and you expected them to remain steady, then there's probably something you should look at. If you expected them, you know to increase over time and they're declining. Then that's also something you should look at, but they they all require interpretation, and this is this is the this is the tricky bit, and this is the part that you know everybody wants just like.

B

I don't know the the canonical list of metrics that I that I can use, but, depending on the just depends on the project.

A

B

And the size of the project and what you're trying to achieve as a project and where you are.

A

B

A

B

Like you know, as a sandbox project would expect to have very, I would I would look at very different things than I would look at a kubernetes, because it's a different just a different phase of the life cycle of the project.

B

This doesn't help you any.

A

B

But there are probably some common things that we could probably summarize this in a way that you know if you have to measure five things. These are. These are a good place to start and what they might mean, and we could certainly do something around that.

A

And actually yeah, I'm seeing more little things. I want to add to a lot of these, like, for example, for a project that has an original sponsoring company.

A

um It really is important, assuming that the project wants to broaden their contributor base, which presumably they do if they join the cncf.

B

A

um It's critical to actually kind of split a lot of these contributor stats out into sponsoring company non-sponsoring company the which dev sets doesn't really help you do, which is something to actually think about um knowing the back end. That would be hard.

A

A

A

You know and sort of other things like some of these things don't come from dev stats, but one of the things that somebody can manually keep track of is missing, expected release dates because one of the problems I see projects getting into is is each release is later than the previous release, and that is caused by a variety of systemic problems.

A

B

Well, let me know if you want some help on the on the metrics on picking the ones that might make sense and whether we might do it do we want to do it. Is this part of this dock or is a separate companion? Dock.

A

No, I I think part of this stack makes sense. Let's just build it.

B

B

I also just while we were talking, I pinged carolyn and asked her to have a look at that pull request.

A

Okay, the um oh did she speak up. No yeah, no.

B

A

B

B

Well, that's it! Oh yep, happily take the extra time in my evening back.

A

Yeah, like usual yeah, like usual, hey.

B

I sent you that question about that fork in chat the other day. If you get a chance and can introduce me to somebody that would be cool.

A

B

One that you don't want to think about.

A

Oh right, yeah, I need to find out who the heck is dealing with that so yeah um I will ask yeah, I I don't honestly know. I know that we use it for a couple of things, but I don't know who's in charge.

B

Okay, yeah, I think it's gonna be somebody to talk to that'd, be cool.

A

So yeah, okay, all.

B

Right take care have a good evening. Bye, bye,.