Cloud Native Computing Foundation TAG App Delivery / Bi-Weekly, 21 Jul 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CNCF TAG App Delivery 2021-07-21

Description

CNCF TAG App Delivery 2021-07-21

A

A

B

Folks, I just I didn't I just want you to keep going. I got to drop off for the next little while because I'm going to a a sig meeting, so we'll see you soon.

A

B

Okay I'll see you soon.

A

Battery 80 connected to josh xps.

B

Hello, oh, did you guys hear me this whole time.

A

C

B

You've heard me getting off my other meeting.

A

A

A

Just giving a few more minutes for people to join.

A

A

D

We gave about five minutes for people to join hello. Everyone welcome to one more um tag, app delivery meeting. I don't have slides today, so we can follow through the document. The agenda is pretty short today. So if anyone has an opportunity to say something after we go through the topics you can, I think so, should I share my screen with the dog after that.

D

So today, 21st um first item in the legend is the operator white paper release, so it happened on friday, the uh the article the article sent by like posted on the cncf website uh there are, it's still open like of course it's like a first version, let's say, but if people feel strongly um that there's some some change that should be made, please make a pr, uh I'm not sure. If tom is there thomas do you have anything to add. Maybe.

E

I'm sorry I joined a few seconds ago.

E

So I think you you are talking about the operator white paper at the moment, um so the idea was we have. We are out with uh with the version one now um and if someone of you finds out that there's something missing or we need another, we need, we need another sections, um just feel free to reach out and open prs and so on.

E

um As the operator working group itself is about to be dissolved at the moment, um so we will um continue the work on the white paper in the in the gap delivery itself. um So if someone of you has ideas to improve the white paper, additions just feel free to reach out and open people.

A

D

Thank you and thank you, everyone who contributed uh during all this time. It's been a long journey, but we are very happy so, thanks for all the collaboration, reviews, etc,.

D

Okay, move to the next one proposal of the cooperative delivery working group, yeah updates.

A

Maybe alex.

F

Yeah, so um the primary update is that it's ready for presentation now to the toc. um I was discussing with some of the other chairs here. um How we do that and I believe it is opening up an mr sorry, a pr in the toc repository asking for a slot to present and so I'll do that this week.

D

Well, apart from these, there was nothing on the agenda um and I don't have any more updates. uh Does anyone have any update or some question or comments.

D

So I believe we will have a shorter meeting this time.

G

I have a question about something that came up uh in the last meeting the last meeting. We talked about potato head proposal for sandbox, I'm wondering what we want to achieve by going to sandbox.

E

um We would like to get our contributors to the potato head, but to be uh to be honest, I or I also don't know really much about the state of the potato head in the sandbox at the moment.

G

I just wonder if that, because we're not going to see it really like how we're going to measure real adoption of the potato head in terms of gradually graduating it between steps, I guess that that's the main question there. If we're not going to be stuck in sandbox, because you can't get adoption for something that is not a service or in like real world application.

G

H

I can chime in and give you a little bit of context. um There is a recording of this. We we had a sandbox review in the toc. I don't know a week or two ago, something like that. We didn't get through all of the applications. There's a lot of them these days, um but we did discuss potato head and the general feeling was that it's a different kind of a project.

H

It's not something that really would what what would it mean for it to get to incubation or get to graduation, and so um the general feeling was that it's a very valuable effort. Everybody in the toc likes the effort, but the general sense was it's probably not a sandbox project because of exactly what you just talked about, which is it's not something that people would embed in a system?

H

What does maturity mean for it? When is it complete? It's not it's a sample application. It's it's an ongoing! It's a it's a valuable artifact, but probably not really. What's that.

B

It's more content than yeah.

H

This working group can continue to care for absolutely I mean this. uh This tag.

G

Yeah, I think, given what you're saying- and uh it might be- that we as a tag need to understand how we want to use the wider cncf to put more uh to make it more visible to make it more wider. uh So end user will know that there's potato head. They can look at that.

G

Maybe adding that to some proposal. So some areas that project can say in their proposal to send. What's going to give incubation, we in we integrated with the potato head, and everyone can see that example. There.

H

B

Yep, if, if there was like a sandbox playground project, maybe we could combine that with this meaning. Is there anyone trying to run like a kubernetes sandbox on a browser or something and say, hey, we'll use potato ad as your tutorial source.

A

G

That's all all for me on that topic.

D

I'm just adding some notes because uh how we would um like potentially did you say like so I understood you well like potentially suggest it as a a way for projects that are looking to graduate to have a sample application added potato head as a step. It's like something.

G

Sounds like from what crawl cornelia said: sorry that the tlc would not uh add it as a sandbox incubation project because it doesn't fit the process. We have right now for that, but it might be the places that we can use the toc processes to have more people integrated things to the potato head, and we might need to understand how the cncf can help us put it uh wide in wider audience. So more people will know, there's a project like that and they can use it.

G

That making sense cornelia.

A

B

You know it knowed, that's just thinking about they. There was a project called canary in the gold mine. Can I hear the coal mine it should have been called, but the idea was that it was like a test harness where you it's still there. I think you can right put your module in and it will make sure that it works with the latest version of node or something like that. Maybe that's like it's. Maybe that's what this kind of evolves towards like if you have a new deployment mechanism.

B

Add it to this.

B

This is our standard test, harness.

G

Yeah yeah. I think that that will be interesting to work with the tlc to understand, if that's something that can be added to some procedures as a way to demonstrate the real use cases and help adoption rates.

A

E

Yes, so this was one of the initial ideas of the potato head to demonstrate: lots of deployment mechanisms and other things, uh ems for the offer, the cooperative delivery working group and so on. We will also use potato head as a demonstration application and therefore I think it might be valuable and one idea behind the behind the sandbox application of the potato head was to make it more or less.

E

Let's say we: we wanted to provide a way where we can say. Yes, we have a sample application provided by the cncf.

E

Which everyone can use to either demonstrate capabilities or otherwise um have have a simple way to deploy a sample sample application, which is not really there at the moment. So this was one of the ideas we had with with the sandbox application, um I'm pretty sure. Even if we get to sandbox at one time at a time we won't, uh we won't graduate so fast.

E

um But um yes, I I think it would be valuable also for the scene save and for us.

G

So I think we can it still sound like we can put two kpis around a potato head. One is to integrate more cncf projects inside of inside it, like I know, uh alias, did it before just approaching them and asking them to add them. So that's one and a second, let's, let's maybe a approach, more tags that might want to demonstrate security or other attacks that just want to demonstrate their capabilities with that potato head. Making that wider tag, effort.

E

Yes, it could also be a perfect sample app in the future for for service measures and so on also.

A

E

Their capabilities, so the possibilities are endless.

D

Right, I didn't understand this like uh for the um in some people's use, like it's not a suitable sandbox project, and then thomas is saying that it might be still like where, where is it hanging out? What's the status of of the of the projecting with the application where.

H

Yes, so um it has not been accepted as a sandbox project and at the moment I don't think will be um I it's.

H

It's always possible that there could be a discussion, but at the moment I did we outright reject it as a sandbox project. Or did we agree to circle back with the tag? I don't remember, um but it has not been accepted as the sandbox project.

G

I just think that if we yeah sorry.

D

So it did not, did it not fulfill some specific.

D

Requisite that is only a documentation. Yes, so.

A

H

I don't know if it's you know to the letter, something that was a lack of fulfillment. I mean, I think the application was fine, but there are some certain things where um there is an expectation when something becomes a sandbox project that it's the first step. It doesn't mean that everything goes to incubation, but there's a vision of okay. This is the first step and then it will get to an incubated status and then it will get to a graduated status, and that is something that we struggled with.

H

We were like what would it mean to have a graduated potato head product project, and so it was more of a is this the type of project that is suitable for, because it's not just sandbox sandbox, fits in the context of this larger governance process, which includes incubation and graduation, and that's where there was. You know we're not sure what this means for this type of project.

H

So it could could mean a number of things. It could mean that this type of project is not suitable for that train, that that sandbox incubation graduation um or it could mean that we need to adjust our expectations, but given the expectations of the current governance process, around projects like that, it didn't meet that criteria of we didn't. We couldn't imagine what it would mean for this to even be an incubation project.

H

So if there's nothing wrong with having experimental work and having you know, things like this, it's just not at the moment wasn't clear that it fits kind of the mold of what these types of projects that go through these stages um are envisioned to be.

G

So I think, just just a thought here we can say that it graduates between stuff, so the adoption is of it- is a project that uh uses it to accept. To put their example there and when we add more project, more sincere projects, we can graduate in steps, but that's again only a thought around that, and it might be that the tag doesn't really need it in sandbox. We need the exposure that sandbox give us yeah and yeah. We might want to do that in other ways.

H

Yep yeah, which gets back to the question of what was the goal of getting it into sandbox in the first place. If it was exposure, then yeah, as you just said,.

A

D

I'll put some notes here: if some people yeah.

A

D

Cornelia can take a look, maybe uh after here like asynchronously, maybe inform if there's something missing. Sorry, hong kong.

I

I

Can you hear me great uh so sharing some thoughts on the on the potato project, like from my experience talking to users like uh most of time they care about, like end-to-end, app delivery experience instead of focusing on like yeah, I have a potato app and I have it running up like people want to see like how I can build the app I can like do a canoe roll out and have a actually something running out and have all the all the matrix showing up so so so I think, like probably, we can extend the project more like to be a like a an end to an epidural experience with like the the matrix rules and all these kind of things.

I

So they're like yeah yeah, you can have 10 different, like service measures, uh implementations right but like they all, but instead of like saying yeah, you have it. You have two replicas running up like we can instead important to be more like a end-to-end epic livery like uh pipeline and have the headlight conformance testing like that. Showing up this thing is running great and also like demonstrating your features into the pipeline.

I

That would be a a more like end-to-end nice experience.

A

I

D

So you're suggesting creating uh like adding enhancing the temporary head to have like a.

I

End-To-End app delivery things right.

I

Currently, it's just like demonstrating like a and an app running up, but that that would not uh I've. Just just from my experience talking to users, they want to see like more uh something something that's more closer to the user, to developers where what they do is actually have an app have the code and and then they just want to deploy it and have something come running up and serving traffic and everything looks great.

D

The only concern with that that I could think now is that we would have to be opinionated about the way to deliver such thing. This end-to-end experience. I wonder you know there are like several ways in the landscape to deliver things, so it would maybe.

A

D

Get it to another.

I

Sdsc, but just like I said yeah you can, you can have 10 different service meshes like implement different features in between, but like yeah you can. You can swap out, for example, for for initially like we can you've built it using like e-steel, but if you want to swap it to some other service mesh provider, you can just you can swap it and plug in your own uh features and demonstrate it that way but like to users, it's more.

I

You know like more more end to end and more complete and and and that that would be a better like demo.

D

So, besides the demo applications, we would have to have another section with like demo the delivery with this way of delivering with that way of delivering that so, but.

I

That that's also inside our tax scope right. We keep not talking about app delivery, not just like um running up a an app right, not just that. If you I want to an app, I could just use a container to run it up or serverless platform to run up right here.

D

Sure yeah, that would be like a growth of the project as well, adding ways to deliver and trend experience. Okay,.

A

Sounds like a good idea.

D

Would that be something that the uh working group cooperative delivery would like uh end up collaborating? I we think like, because we we are looking into ways of um yeah.

I

Gaps right yeah. That would be great suggestion too, like uh we can actually, along with the epithelial pyline like there's, not only one web service app, but also like some data back-ends, um just like putting up as well right. So in the storage.

E

And there's also, there is already a branch open for uh from marty service implementation of the potato head. Currently, we are playing around with taste and tone, but this it will definitely get more than one than one service in the future and also we we plan to have stateful workloads in the future.

A

C

Was going sorry, go ahead.

A

H

We might have enough people to uh to to do hands. um So quick comment, hong chow, I think that's a great idea and while you were talking, I went over into the potato head repository to see about a road map and the road map is pretty thin.

H

um There's not much captured there, but I wonder if you can capture your idea in an issue and then um you know that is kind of orthogonal to the whole, creating a roadmap for potato head. But I would. I would suggest that you maybe capture that in an issue in the repository good suggestion.

G

Yeah sure do we have discussions enabled in that repository.

H

G

Your fancy tools.

I

Sure yeah no problem.

D

um So yeah we don't have anything else in the agenda. Thank you for uh yeah, like bringing this up. The potato stuff was very interesting and I think there's a few notes and actions captured on the meeting agenda with some notes. um Yeah, please I mean I didn't know about the discussions on github I'll. Take a look because I haven't looked at, but yeah it will be good. We can start using that.

D

Does anyone have anything else to say before we end the meeting.

C

I did have one question about the cooperative delivery working rope, so one of the goals is to provide some examples on the potato head project, but my question is: isn't it too simple? Maybe for for the kind of uh you know the the approach that we want to have the breaching the gap between infrastructure as code and application delivery?

C

Is there any actual need for infrastructure beside the underlying kubernetes cluster for the potato head that we can leverage to create these examples?.

F

I think that's a really good question um and yeah. We would probably need to grow the complexity of potato head by a degree, and we can do that, especially when we think about things like you know: state management, data stores and uh you know ingress at the infrastructure level. So that's a conversation and perhaps something we can tie into that potato head road map and perhaps it could even be fuel for that road map is to make it relevant to the work of the cooperative delivery group, but yeah absolutely spot on for calling that out.

F

Like current in its current state, it is fairly simplistic but.

G

Yeah and if you have a use case that is more complicated or requiring infrastructure, that will be a good start to adding that and like let's spot the demand when we needed to not try to anticipate what people were needed from that potato head. uh That will also help to spot project or working group that are working on something that no one else is working, because if no one else had had that, that's that's a green field. To do.

A

Okay, thank you.

D

Okay, um we'll have a few more minutes back of your day. uh Thank you. So much for attending and uh giving your opinions and comments uh have a nice week. Everybody.

A