Cloud Native Computing Foundation App Delivery Special Interest Group, 1 Apr 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CNCF SIG App Delivery 2020-04-01

Description

CNCF SIG App Delivery 2020-04-01

A

A

Hello, hi, everyone welcome to the seeker clear meeting and my name is Lee and I'm co-chair of the seek and very nice to meet you guys and just a reminder that the whole meeting process will be recorded so place of greeting there. Okay, so we do have some simple, very important items in today's agenda and, first of all, we like to know the recent update from the air-gapped working group because I remember last time the air-gap working group talked about the partner and we also talked about that. We need to approach to the community.

A

To get was the recent practices of the air-gap delivery in the current ecosystem? So I'm not sure the team years from air-gap working group, Edina, meeting yep.

B

I'm here, okay.

A

B

So I guess our update is we had one meeting so far, one official meeting after many many doodles figuring it out and we've to start a charter and a couple of documents that we're building that are linked in the agenda for this meeting that you can take a look at the first two items we are targeting is collecting our users, how they are currently installing kubernetes and aircraft environments and we're going to collect those cases and develop best practices document and we're doing the same thing for chain-of-custody.

B

That's one of the biggest concerns that people have with air-gapped is how do they consume upstream artifacts and maintain chain of custody when they're in an air-gapped, so we're starting another document for that and collecting what people are doing and how they're doing it and help to come up with best practices? For that, so that's kind of the first two things we're focusing on either of those things interest you you can even fill out the document or you can come to our next meeting, which will be this Friday at.

B

Ten o'clock a.m. Pacific and that's kind of our update, happily answer any questions.

A

Come in contact with they said to you know to to eradicate all these documents.

C

Yeah at some point, I think we need a way where we can store them. We could actually make a page on the github repo, but we still working draft documents. If we wanted to.

A

Yeah you just found this new house.

C

That's not my new house I wish this was my house. This is actually a 25 million dollar today.

C

D

C

Right now everything is fake, so that's the exit. Okay,.

C

Yeah, so we can actually have working documents. We don't put a lot on github I was just looking at the github repository. There's nothing really there and we could make folders for working group, so they can post all their documents there. So I think that would be a good idea. Yeah.

A

C

I can do that to create the folders and everything, and then you can file at PR against it. For those who don't know where it is, it is actually here, but that's good, because that was one of the comments in an earlier discussion that we don't have a lot of things online. Actually, so this is already in case you didn't know, we have a github repo, where we can store stuff.

C

So we create one subgroup I'll take. This is an action item and yeah expect it within the next couple of days to be there.

A

All right so I think we can move to next topic, which is about the operative working group. Naturally, culture is here.

A

Seems not yeah.

C

I can also give an update because I was part of it in the last time so yeah. We have two volunteers now who want to chair that working group, as mentioned in the talk as well. Let me post each charted here for those of you who haven't seen it sorry too many documents open right now.

C

Okay, so I can let just give me a second.

C

Say that the list is impressively long of people who are there so Jared Marc volunteered to share the working group we have already. You should have already received and meeting invite I think it's every second Thursday that we set up that meeting and the first work item for the working group will be to work on the operator definition that we started a while back and then type into other topics. I think there's some additional comments that Matt posted in there.

C

That should also address in the next meeting and overall we should bring it up in the next update to the tea. You see as well that this one is forming right now.

C

So not not a lot of work done yet, but there are scheduled meetings on the calendar. We have people who take care of the work and it's going to be done and first topic definition. Then further topic should usually be also collecting the best practices. There was one of the big discussions.

C

There are lots of best practices out there over how to write and start your operators and also helping some other people on developing operators in case need, support and then eventually moving into more like in interoperability type of discussion and also relate already interact with other projects. We already had this discussion between operators and air-gap. Even there were some issues there, and also with some other company disrelated working groups that there were certain requirements and that the working group can bring up these topics, and we also discussed and the flow should be.

C

The working group brings it up to the sake the Signum can help to engage with with the other projects, but we are inevitably removed as non-course is to be not going to obviously write our own s, the case for building operators. There is already quite some out there. We don't want to create yet another one and per se. It's also not the goal to recommend a new project, step deceit that should be added to the CNC F for the time being and also excluded everything that's non. Kuva needs related, operate device.

C

B

C

Pretty much the work that was done there I think more impressive things, I think the list of people who are interested is longer than one single page, so I expect, hopefully things to move forward rather quickly. It's.

E

A Maggie the ready huh go ahead. Man I just wanted to note one thing: I think you said the meeting was on Thursdays according to the CN CF calendar and it's listed on there it's on every other Wednesday. It.

C

Can be as well and what I might be wrong. It is on the calendar, so I might be just wrong about that. One.

E

The first one was on Wednesday is what it says: I missed.

C

E

C

You're right, yes, the next one is your right is next Wednesday and it's 6 p.m. GMT one and.

E

C

E

Gap meet yeah this and the air gap meetings are on the CNC F calendar or the TOC. So if you go to the TOC github repo, just in the readme there's a calendar link there. If you go there, it has every meeting, including these working group meetings on there. If you're trying to find out when something is alright, that's it.

C

Yeah but thanks for pointing it out, but yeah, we have all of these schedules and let's wait for the next update in in two weeks from now. Don't expect too much progress over the next two weeks, but let's see how far we can go. I.

A

Also noticed that in the definition of operator and work we actually did, automation was a difference between controller and operator that.

C

Was one of the big discussions when the question came up? What's the difference between a control and operator and the arrow topic was can something being operator if it's not using a C or D okay, this particular definition. It does not, but we had a discussion about flux, flux uses a git repo.

C

So it's like I was like a wood flux, count as an operator or doesn't it count as an operator so but I think that discussion will be interesting. What's the controller and what's a what's a C or D.

C

F

C

That's pretty much it much the update.

A

Yeah I think the operator working group need to raise a wire discussion about these fundamental issues and they want to collect something back from the community about okay. What is your impression of open already, because I do talk, we say whoa guys from the community, and the feedback is more like the operator is for you to install a software instead of all cranium started here and I. Think that is one of the common misunderstanding about the open era. We are talking of the automation the day, 2 and phase 3 operations.

A

It's still consider the autumn scope of the current operator in limitations, I.

C

Think what we should do going forward here once we have a first definition. We need to get back to the user and ask if you see for the mimic I think they have is wrong opinion and what they want to see in there right.

A

Okay, so let's finish race topic and go to the third one is discussion actually about Harvard. Graduation really- and we talked I actually talked- means the Tuesday, but Eastburn and I. Think it's more like a collaboration review from multiple states. Consensus, not sure if there anyone from linear to give us some more feedback, yeah, okay,.

G

I'm right here this is Michael I'm, one of the main tenets of harbor and I'm, the one that added it to the agenda for today, essentially Harbor has been, has been up for graduation for the last six months. We've seen C F and we had to go through a set of reviews by seek run time, which gave us thumbs-up six storage, which had one concern but and everything was good with with harbor and then were undergoing review. Six security now Quinton had said that maybe sig, ABS or sick up delivery should also review.

G

Harbor I personally didn't see a connection there, but you know, following with the sin CF guidelines, so I wanted to talk with you guys and see what kind of review you can do or with what context in mind, since we already had three other six review Harbor.

G

The ultimate goal here is, for your sake, to either give us a thumb, BA, abs or thumbs down, or something in between on on our bit towards graduation, with the focus of what you intend to review in the in the meeting notes: I added a couple of items like our PR that we'll have for graduation see gobs review ticket that believe it was a record that I created it he's from sick run time and then also the due diligence document, which is some. If you are, if you want to review harbor.

G

This is basically how you get started in it's a fairly comprehensive if document that talks about the cincy of graduation requirements and how Harbour meets them in each one of the different categories.

G

So what are the thoughts from this? Second and another? Mod has gone through this process with helm just very recently. So if you have any thoughts on that.

E

Can I can I first jump in and just say something real quick uh as somebody who's gone through the process of putting this together, it is incredibly laborious to collect all of this.

E

Are pretty small going? Can you hear me.

A

E

Benefits are pretty small, going from incubation to graduation as far as you as a project have and the amount of work you go through.

E

I want to commend you for for putting this time, and you did more than I did for helm, especially going between all these sakes, so I just want to call it and give you props for for having the energy and the stamina to go through this for so long and to do so much work and write this up stuff up so clearly like your due diligence, is what 29 30 pages- that's incredible and so props to you for for going through. All of this, that's all I wanted to say yeah.

G

Thank you much appreciate it. Yeah.

C

From from our side, maybe what we can do is the sake so I don't actually see any issues that we would be bringing up. What we can do is have to share the details. Gents document we can also provide, usually, at the very end, be provided to pass this recommendation section usually- and just let give us maybe a bit of time to read it through whether we can make the recommendation from our side as well. Yeah.

A

Yeah I actually will just do the documentation before and also making person. This is Henry.

F

A

Veneer about this project, so I do have several questions about harbor. If you guys can provide information on the documentation to do so. Its first Louis I would like to see what is the current artifacts in harbour. Example: are there or dock images, or there are some other kinds of OCR artifacts story in hopper, I.

G

Can answer that so so today, Hart bore our latest release is 1 to 10. It only supports hum charts some charm to Museum, and then it also supports continued images. Our release of harbor to the dough that ships this month about three weeks away from the release is fully OCI compliance or any OCI compliant file can be managed in harbor. That means you can push it. You can pull it. You can replicate it because harbor has a very extensive replication engine.

G

You can enforce quota management, you can enforce retention policies, you can enforce our bugs or all the features of harbor apply to full OCI compliance, and that's, we actually have been demoing that for the last few weeks, also our scanning capabilities apply to OCI compliant. So, for example, if you have a Sina bundle, that's a thick bundle with 4 or 5 images in it. We can open it up and scan the individual images and give you a compliance report on them. Yeah.

A

I think that is the most important thing for the egapp delivery is carry about, because harbor now can be used as a distribution center of a software animation. It can be any kind of OCI compatible, artifact, storing harbor and you can take to peace takbir from that kind of on that source. So I think it's important to added these part information in the editing, documentation and I think the co-chairs of seek actually really also pay attention to that part to see.

A

Okay, if harbor is a good candidate or is there anything to need to be improved.

F

A

To acting as a software distribution source like dhaka, harbor right, so I think that is one thing I want to you know and to check out all the documentation, and the second thing is about the there's scalability of harbor, so I'm, not very sure about how many you know knows or how many I know the matrix you in the harbor, so I do yeah I do so. There is Coverity part in the documentation. So have you ever considered about this clip scalability? Maybe has some.

A

You know metrics related to software distribution, because today, I see what I say is not about okay, how many nodes, how many resources matrix the the Harvard you things that have ever considered about you know. Maybe you can have some part of the scalability from the applications of also to give software distribution perspective.

G

Yeah so so so, basically do scale testing from a few different angles right. So one of it is from the end-user think the developer. How.

F

G

Concurrent push and pull operations, we can do how many images we can manage. You can emit, manage 100,000 images and we actually have customers that have terabytes and terabytes of storage on harbor with you know, thousands of images and that works fairly well from if I understand your question correctly.

G

You're also asking you know: harbor convene a hub-and-spoke model right where harbor you have a single harbor instance in the middle which can enforce your policy around scanning and compliance, and then you can replicate those images all over the world from the central location where you have clusters on the edge. Just like folks, wanna put a kubernetes cluster on the edge.

G

You might have a harbor instance on that or if you have assets in a small medium data center all over the world, you wanna replicate themselves central harbor does all the computer on scanning and compliance, and then it populates the images to other harbor instances or even other registries, so that they can be utilized by developers.

G

Harbour is horizontally scalable. Our services are stateless. So have absolutely no no problem scaling to this, and we have, even though traditionally most of our customers only have two or three of these stumps that they replicate between.

G

We have tested with combinations of many many replication column, replication workers that replicates images all over the world, but typically customers only have two or three, so they have a couple of data centers that use them for jury, dun, dun see and they replicate their images across that who haven't seen people start putting kubernetes clusters on the edge, yet at high volume, where they also put harbor on the edge okay.

G

One thing I wanna mention that, since we are talking about the edge, are to do one release of harbor, which we've already started working on now it will probably ship prior to Cuba con, if not much earlier, we're actually gonna enable proxy caching capabilities in harbor, as well as p2p distribution. So that's specifically targeting the edge scenarios. So if you are putting a kubernetes cluster on the edge where that's a 5g, that's with 5g that's going to become more and more the norm, then will enable you to put harbor.

G

There will act as a hot proxy cache for the images that you care the most and then it will push them in from the main data center.

F

G

Then we can also use p2p like dragonfly and kraken to push the images there in a more efficient fashion. I see.

A

So yeah I think my question about I. Think you can, you know, do some more calculation about to make matrix more and the user facing, so the developer will know. Okay, I have maybe 1,000 node cluster I have 1 million applications to distribute, then what kind of harbor architecture on e who use what? What kind of how many nodes for harbor on deploy? So it's more like any other facing matrix, I hope to see in the documentation.

A

So this is that in the second question, its third question we actually you mentioned I I- also hope to see how harbor cooperative is the existing software distribution technology is like dragonfly right. Like you know, goober also has seen in a project which came in a bootstrap. The image distribution, rhino life I thought I noticed that this part missing from harbor. So what would be the common status of that integration? I.

G

Just I just answered that right so I said.

A

G

To that, one will have integrations both Kraken and dragonfly, so that cracklins from uber dragonfly from Alibaba, and we will enable p2p distribution of those images. A lot of that work is not dependent just on harbor right. It's also the communities from dragonfly as well as Kraken, to enable that integration, but we're already working on it. I think we have have couple of POCs already that that you.

F

G

Of that work and then specifically for dragonfly, we've shown that them already and then the proxy cache capabilities and I think Matt is raising his hand.

E

Yeah yeah I, so we're gonna hear some of this. What I'm, also wondering is, is how much of this is part of the graduation review or kind of just a broader discussion right. So I I heard something that was a suggestion for docs to do better and improvement on Docs, and now every project can have better Doc's every single project out there is this something that should be filed as an issue or is it something that should be part of a graduation review, that's kind of where I'm getting to to this whole thing?

E

Is it a mature enough project that it meets the graduation criteria, or are these incremental suggestions for improvements which every project is going to continue to have and and I just want to kind of keep that in mind so a project, that's under massive scrutiny, I mean what they're going through for SIG's for reviews and lots of opinions are coming around. How can we kind of keep the graduation criteria first and the other things we kind of note that their suggestions I, should go in issues but they're, not a graduation blocker for them.

E

That's the only thing I wanted to bring up in mind because I thought, like the doc suggestion, it's a fantastic suggestion to file as an issue over there, but it's not a graduation blocker kind of suggestion. That was my only my only thought. I wanted to jump in with all right, I'll be quite again.

A

Yeah I know see: there's any graduation blocker we're talking about. Then you need to create even include documentation to like the seek to know. Okay, what kind of all the keidel functionalities your project related to the cig? So it's better for it's better to have that can include service. It can kill the intruder in.

F

A

It's kind of that kind of discussion, but that also possible that these part of information can be a blocker for the graduation. So just keep that in mind. Okay, this is actually three questions. I read through the documentation, I hope that you know hubber folks, can add more information about relationship with the software distribution, because I personally can see the relationship. That's why I think khuddam is recommending recommended and you guys may want to go to a posh release a little bit yeah.

G

Yeah and like I mentioned, you know, the dragon flying cracking proxy cash will be well documented, wouldn't 2.1 chips so that that's guarantee.

D

C

Think if you find an adding our statement to the review, but it's a Matt's point here- things might not be part of the review, but that kind of of the graduation review but I think it's still an opportunity to get some feedback from the SIG's which we can provide on certain topics, and this case they are all covered and the questions came up. So please also take it. Take it as inputs that that might make sense, for some of these things are not just as a graduation blocker per se.

C

So if you take the time to discuss topics, I think it makes sense for all of us to to share feedback as well yeah. They give us time- and we can't have like generally give us time if you can, after that, if you know note down there, but I'm, not sure where it is like massive blockers, but obviously ACI compatibility was one mentioned by Harry. Did so.

G

Yeah one of the things I've seen work with other six is assigning an owner from the sig, that's responsible for kind of shepherding the review and gathering feedback and providing it without that person usually I've seen that the process takes time like a couple of months. Sometimes so is there someone from C cup delivery that wants to that doesn't have a conflict of interest. So, for example, cannot be Brandeis who's, not here now today, but who can?

G

Who can basically own this and say I will share the feedback, go through the review and give the recommendation.

A

I think I, just as I me and the contact person for harbor project, if you haven't, have not approached anyone before. If you have that you can just continue approaches. If you did not, you can just add I me as your contact person or the harbour project review and, of course you can get all the feedback from any other community, but I don't think that's related to the graduation process.

G

It's not up to me to assign um if the sick up delivered, if you guys are okay with that, that's that's the the seek the decides that not me! You.

C

Have a two-thirds majority here that, yes, yes,.

A

You can you can always ping me on this lag. So I can be the contact person if there's or if you need not anyone yeah.

G

Sounds good I'm happy and are you thinking what kind of time frame just so I can also set the expectations on our end? Is it like a couple of weeks, I.

A

May have there were other questions about the detail body.

G

Absolutely maybe.

A

I think we can I think we. Hopefully you can finish all of the work in one week. It should be quick. Thank you.

G

I'll Pinkins like and I'll, be able to answer all your questions. No.

A

Problem question.

F

Are you doing the review in Google Docs or over in github, just want to be able to know where to track this.

C

So sorry, because for us, dogs were commenced and.

C

G

Doctor is a Google Doc, um but ultimately what has been asked is that you put your feedback and github issue. That's the github issue, that's linked in the meeting notes as well. That's.

F

Fine I just want to know where to be able to track for progress and if I need to go hunt through Google Docs, that's fine if I need to go hunt through get to it. That's also fine.

F

Let me know thank you.

C

Okay, looking at the agenda.

C

Who put artifact up on this one I did. This is Matt.

E

So I wanted to talk about the artifact hub because it is now being proposed for sandbox and the issue had it assigned to say gap, delivery.

C

That's the nice one yeah surprise.

E

C

E

So I'm happy to walk through and talk about it. However, y'all want to proceed, but I did want to put that up. There.

C

It's fine, so I think yeah. We can store it with with the with the review documented artifact type I think there are some questions open around artifact hub.

E

What questions we should get them documented on the pole, request or the issue so that way they can be answered for everybody, yeah.

C

My first question honestly about artifact tab is even even for a sandbox who will be the long-term maintainer of this. This is now CNC term long term maintains because there's no like like, unlike other projects, the CNC FF, was created our efficiency F and put out the contractors that used to work on it and what's the long-term maintenance strategy. That would be the first one for me. Well,.

E

I guess the first thing is we're gonna work to try to engage more of a community around this. So it's not just long-term contractors, but we have not worked out all of those logistics on that yet and I don't know for a sandbox project if that's a requirement, but that's one of the things that we're gonna look to to engage more folks you want to, but it may be a mix of contractors and maybe a mix of other people. I, don't know, we've got the CN.

E

Cf has contractors who work on lots of code bases for it. um More than I I originally knew that's kind of a Linux Foundation thing. They do regularly have lots of contractors working on things even over in, like the open containers initiative like I, know somebody who's doing work over there, so we hope to have a mix long term right now. The project is just being heavily sprinted out by a couple of contractors and they are doing a wonderful job, but in the near future we are planning on going into a wider release.

E

We're not exactly sure when that is more when it's not so pre-alpha, it's probably more alpha or beta right when you think of like a Google beta, it's rough around the edges, but it's ready to go we'll, go to a wider audience and we'll start looking for more more basically web developers to come. Do that which is a different sort of thing for the CNC F, so we probably won't be picking from the normal ciencia folks, you do infrastructure things, but other people at those organizations and and stuff like that to engage on it.

E

C

E

C

So I mean obviously we can go deep on a review. I think it would still be good to if you go out there and could engage or recruit some people from CNCs member organizations that are willing to actively contribute, especially given that some of the member organizations initially so it's more less as a competing projects to what they have put into the CNC F and they're working on. No, it's not necessarily a requirement, but still for sandbox. It must be an ensure that that, like long-term interest by somebody working on it yeah and we're gonna.

E

C

Out there and find people who are actively willing to contribute, especially given all the TOC related discussions that were held around it's just the recommendation to you guys, yeah.

E

Yeah- and that is in the plan- a meeting with the end user community later this week, because we're not just going so one of the big benefits here is for the end user community for people to be able to discover this, and so we're actually I've met with some of the end users and have met with the end user community, where I've got nothing but positive feedback on doing something like this, and so that's one of the areas where we will go. Look for folks to engage on right now.

E

You know it's again it's in its early stages, so we are doing this but we're being slow about it, especially as we spike things out and make lots of changes. But it is in the in the planning.

H

I'm curious at this juncture who the we is that's working on this because it seems to be a party of two or three people, the contractors and yourself, and so you know one of the requirements for sandbox or incubation, or anything else like that is that there are active participants and contributors to the project. So as we go through the review process, I think as Eloise is saying that it'd be good to have more committed people working on it than just yourself. Yeah.

E

And those are not requirements for sandbox for the sandbox level having people from things like multiple organizations are not requirements. It's one of those things, in fact, in the sandbox requirements, it actually talks about things that were commissioned by the CNC F to get going.

E

Those are candidates for experimentation, which is what the sandbox is for, and that's why you can easily archive things out of the sandbox if they don't go well, and so, given the current number of people and structure it does fit per though the sandbox and there's a separate, sandbox talk that just goes through all these things. It does fit some of the criteria in there, which is something that I didn't notice at first other people pointed out to me. It.

H

Still I would like to see more contributors to this project and more folks actually committing to work on it with you. Certainly.

E

It's still not a requirement for sandbox according to the sandbox documentation, and that's one of those things right. I know, I, know people who've done, reviews around sandbox who've been asked to do things that were incubating level things and so for this I'm really focused on the criteria that is sandbox in this case and so well I. We are going to go out and try to find more people. This does currently fit the sandbox requirements, which is what I'm focused on.

C

I also recommend, but also did for other sandbox projects, see how far along they are from occupation related criteria. You don't have to do it I know, but especially with a project like this, and you know they need the background and, like all the concerns around it, I think it would still be good to address these and just don't go by the books here.

C

But we have obviously people here who are concerned about like how this whole thing into life and, as you rather see the one driving this forward, as he also you being responsible for handling concerns by people, because as a community here, we should listen to concerns from other people and try to help to overcome those concerns. So that would be my recommendation here. So as you preparing for sandbox to also address these, these things and our I mean doe.

C

Note it and there's like like a healthy flow of commits from different people, how you engaging the communities, how people can be on board I, guess it's all a list of stuff that still also make makes actually send as well and yeah.

E

C

And I think it's still also good to have maybe a bit more of a deep dive presentation on opera I want to say operator without that's.

F

C

Co-Writing with the naming here a bit more deeper presentation, but you want to go what you want to do like for the year after livery working group, just as a general information as it fits in here and where you want to go, it would just be. My proposal doesn't need to be the next time, but I think it would so be good, because a lot of people have open questions still and to give them an audience to ask these questions so Matt you can think of.

C

Maybe you were together with somebody else, who's working on the project. Doing this. The next meeting with everything after.

E

So what kind of information are you looking for in it? The history, the purpose that kind of information how it's built? What it's doing well like what kinds.

C

Of related, are you looking for out of the MGF Pro yeah designs and how it's related to other CNCs projects and approach its? What your roadmap is, where you want to go, if you need to.

D

C

You have end-users like who's like using it right now would be interested. Maybe yes, as part of that conversation as well. Okay,.

E

Now note, the end-user part is an incubation level requirement as far as those kinds of things go so I'll drum up some things on that. But end-users is an incubation requirement for those not familiar.

C

It is, if you have it, I will recommend it. You know the background office and you notice I I.

E

Do and I know that this was started as an experiment. That's in pre-alpha! That's not targeting lots of end-users at the moment, because it's attempting to get up and going before.

E

We shoot after lots of end-users and fill out some of the rough edges, and so right now we haven't targeted a wide launch yet and that's why it's very much an experiment at this point, and so it's not trying to be an incubation project with lots of users and we will be getting there and in our next meeting, we'll be talking about exactly that to come up with a timetable.

E

But at this point in time we have not targeted a wide launch with lots of users, because at that point you've got to stop doing some things like breaking changes and things like that. It is very much an experiment in the terms of a sandbox experiment, which is one of the things that same doctor projects are for and so that that's, why we're not going to have lots of end users, we're not targeting it we're targeting the experiment phase. So I'll attempt to talk about that portion as much I.

E

Can and we'll see how far we've progressed by the time I come to present on that. Thank.

C

E

Are there any other questions or comments on this? Yes,.

A

So well I'm hearing about you are now targeting for dodge skill user. You mean for now.

E

We do plan on going a large-scale I mean the helm hub goes out large-scale right now.

E

We've got I, can't remember how many organizations have put their things in, and so a large scale of users of people presenting things and large scale of people consuming things through it, both as a hub and it'll, also be packaged for for people to take a look at, but the primary is the public hub at the moment, and so when it goes to large scale, the idea is to be able to serve people around the world and many geographies in that way, and so it's a little different than many of the projects that come through here, which are meant to be primarily self hosted.

E

This is a kind of a different style project, so.

A

So if I understand correctly so you're saying that you want to make the artifact hard and successful and helm hard Oh far.

E

More successful than the hell that's.

A

E

Goal far more successful than the helm hub, Pagani yeah.

D

E

We're gonna list things other than just charts in it. So I was at the opa call earlier this week and we were talking about how do they get their things listed and we talked to you know. It was just a back and forth I presented what we're doing and talked to them about how they would want to proceed, and they very much want to have policies listed and discoverable. So it's going to be more than just things like charts and applications, we're actually looking. How do they do?

E

Policies and they've been talking about sharing policies and they've got a couple of different ways, and so they're gonna come back to us because we gave them options. We will do everything in the artefact of ourselves. You can come up with things you want. They're gonna come back up to us with what they would like to do and they're working through that now and so you'll be able to discover, hopefully, policies in there once they work out what they want to go about.

E

Doing that and you could have different organizations sharing their policies even as starting points and things like that, and so that's one of the bits that we're doing, and so somebody who's not just looking for a helm shirt. Somebody who's looking for an OPA policy will hopefully be able to go there, search for it, filter on it and discover those I think.

A

That that part also need to be added in the goal or motivation for the project which I think it missing from counterproposal. What is gold missing? What is the goal of the project is count missing for me the proposal proposal. He says what is is hard before you know any one nation that the goal of the project is far okay well,.

E

In in the poor request, the language I have, our notes is the goal of the artifact hub is easy discovery of cloud native artifacts, for example, helm, charts, Falco rules, etc, and to provide information about the artifact so that end users can make decisions about using them. Yeah right and.

A

So so kind of different for multi mission well different from from will teach estimation. So well you just the mission that the goal of the project is far popular. Yeah am.

E

Harm, oh I, I think that that's kind of a personal thing so from the helm community. If the artifact hub takes off and is legitimate, we will be deprecating the helm hub. We've already agreed to support that. If it is able to turn into a real supported thing with use, we will be deprecating the helm hub and pointing people to the artifact hub as the place to do it, and we will work on things like so that helm, client right now can search the helm hub.

E

You can use helm, search hub and then search for something. We will gladly do things like point you over to the artifact hub to search for charts. So that way we don't have to do duplication. Our.

A

Hope is, this can take over our meeting.

A

E

So I'll try to add more of that, like the helm comment is further on in the proposal. That's in there I'll try to find and bubble some more of that up, but that is in if some of this is in the proposal right.

A

And the second thing I thought is missing is what you are looking for. Funsies, yes right, it's a common question: why we want to move already found hard to things yeah. That proposal.

E

And you know it's interesting now that you say that I totally get it because I've told people this in the past and I was going based on the the structure that I was pointed to in the templates and process documentation which doesn't ask for that all right. A lot.

A

So, like you want me, if you don't think you should be part of the template, it's just a question we are asking and we hope that you can added that part being the no.

E

No and I totally agree with that, but I think the interesting thing is in the process: documentation when I was walking through it either I skipped over it or it's not there, which may be why projects aren't, including it. It's not thought of in just that. I need to write that down. No, that's a great question and I will write something up on it. Yeah.

A

Good yeah, this is two questions: I'm, just raising my hey now I'm. Looking at your proposal.

I

Hey this is Rob I had a quick question. I know that we had a call in the Kris Nova organized and she brought some things over to the Falco community and it seems like they're, not in love with this idea and probably won't be participating, I'm just curious what your your thoughts were on their feedback and if you're gonna take me that into consideration. So.

E

We got I got on a Falco call in their last meeting and it's one of the things that we talked about and I think there was some confusion.

E

So when the Falco community was talking about support and what they support, they end up talking about what do they actually contribute people and hours and time and support to and they're stretched, pretty thin right now, they're talking about scaling back some of their installations and some of their support, because they're getting so many questions for things that are outside of Falco so like they have a Falco helm, chart right now and they get questions about helm over there instead of at the helm, project which I totally relate to because in home, I get questions all the time I'm at home project about kubernetes, general things, and so right now, they're actually said they're dealing with that problem, and so they don't want to put ours in and support it and do that kind of work in the shape its in, and that was not something we were actually asking them for, and so we got on.

E

I asked for their thoughts and their feedback. How they want to go about doing it, and so one of the things we're doing is we we are gonna, take their rules that they have listed and display them through there and they've. Given their blessing to do this.

E

They just don't have the time and energy to put work into it and they don't want to put an official seal if we support this, which means we're gonna, put man-hours and our people hours into making sure all the things work, because they're a bit overloaded and I think some of that context was lost, but they are very happy to have things listed and searchable and found they've, given feedback on how we should display things over on the artifact hub feedback.

E

That's actually been rolled into the hub already, since the meeting which was last week, I, think and so we've already rolled out those changes to to their requests and so they've, given positive feedback and haven't been so much against it as they're kind of an wait and see, but they're happy to have their things more discoverable. So.

H

Matt, that is really not the way that that falco call went and which.

E

Falco call I'm sorry this are you talking about day end the call that we had the day before the toc call or the normal standard Falco meeting call the.

H

Normal standard one and to the issue in the comments- and it doesn't look like there's any confusion but and that did look like they did- offer that they should be if, if it is listed at all that they should be unofficial and unsupported by the Falco community. Yeah.

E

Yeah unsupported, as in their their definition of support, as they are putting hours into making things work that that's kind of their their thing. It's not official and and we're not asking to be the official thing. In fact, in their call, they talked about how they're reworking how you install artifacts.

E

So one of the things they're going to do is they're going to stop supporting the Helton chart, because that's turning into a lots of problems and they're, also not going to have an operator or something else, they're moving to lower-level way of doing it, and- and that is still to be determined um and.

E

So I mean I'm, not sure what else to say that they don't want to support it, but their definition of support isn't to give it a thumbs-up or thumbs-down. Their definition of support was to put people hours and time into working on it, which is one of the things I they clarified. For me,.

I

That totally makes sense, so who's going to put the time into like list those things. Is it your working group, it's going to do that, and then is that extended to other types of entities as.

E

Of right now, the artifact hub folks are going to do the best we can to list those and describe them to make them discoverable, and we point people back to the root source on that, because we want them to be discoverable and as they make changes, we will have to adapt.

I

Okay, awesome, yeah I would love to see that process written down just so that we have it so that the next folks that want to do the same thing can reference it yeah.

C

Yeah I mean some issues, even if you want to be pre-alpha and early stage, there's not a single reason right now, if artifact up and when you claim to be alpha, but there isn't even and also release right now. This might also be something you want to look into yeah.

E

And with this, this is one of the things that maybe I'll have a question on.

E

So websites will often in web apps will often work in a little bit different manner than something like shipped software, and so we are in pre-alpha, not alpha, which would be tagged releases at that point and and I expect we'll it will tag it when we go to what we would consider alpha, but when you're deploying something like websites like, if you go to most of the project websites, you're not going to see things like tag releases on their websites or their web apps that they run for their things.

E

And so how does this work differently than the standard ship software mo in this case? As we go through the different steps and and I? Don't know the right answer to it? It's one of those things that is slated for our calls, but that's one of the things that we'll need to work out like when do we tag and say releases and then because then you get in to install instructions when our primary is a web app and and how is that different here, yeah.

C

Because when I look at the documentation of github, it mentions that I should install my own version of the hub.

E

I'll have to go, take a look at that, that's probably for development purposes and on-prem, but our primary target at the moment is getting a working web app up, because the goal is to enable discovery of things on a large scale that are distributed, and we know people also run it on from the software that powers the hell mob started as on-prem software. That was repurposed for the public, and it's just had great success. So we know there's both cases, but instead of the helm hub software, which started on Prem.

E

This is starting more with a more of a public which is actually letting us build out. Certain features like organization and user management, and things like that, because those cases can be a little different when you're on Prem versus at least the unprimed scale that we had in helm hub so and.

C

I mean what our point here is also as we as we move forward on these. He moves for one of the two divisions of the operator for your movement operator. How did you relate procedure that will be also part of the discussion? Oh.

E

Yes, absolutely I'm actually waiting to see what happens with the operator framework and the operator hub. As far as the TOC and vote goes, there have been a couple of different suggestions on the last TOC call and I have no idea where it stands since then, and so I've been kind of in a wait-and-see mode to address that until whatever happens, falls out and and I'm just not privy and haven't, read those conversations and I don't want to assume anything. Yeah.

C

I mean it might also be worse for you still to proactively reaching out to these communities. I mean no, it's not part of the sandbox requirements, but you know that this project has already some history and some very strong emotions to it. So I think it's in your very own interest to you.

E

Absolutely absolutely and because the TOC has been working with the operator framework, folks at least I hope working, because the TOC member took the goal of figuring this out and pushing it through since they took that on I'm hoping they do.

E

It and I really didn't want to step into that middle of that conversation until things were ironed out and I saw the TOC moving forward with whatever they're going to do and I'm just trying to let them work out between the two of them instead of stand-in Diane may be able to give some update on that, but at the moment I'm just trying to let them work out whatever has to happen there and then stick my nose in and start asking questions because I didn't want, interrupts Haidee.

J

From the TOC is on and she was doing the due diligence and then Brendon also was going to look at it. So I, don't I, don't know Katie if you're on and you can glean some so I think brendan is actually the TOC responsible for Brady framework. But at the moment there is a strategy to come well to kind of separate the operator hub from the operator framework and we'll try to push that forward. They says I think they should be communicated last week the end and they should be updated already.

H

We have we've not received that, so we would love to get that officially from you guys as well. Matt you've been invited multiple times to join any of the operator framework, sig meetings so anytime you want to come and reach out you. We have regular cadence of meetings. Oh yes,.

E

Certainly certainly and III plan on it, especially around these topics. uh Quite frankly, my time it's been my problem lately with everything going on and and kids home, and just everything going on. My time has been one of those things that has been a little bit stretched the last several weeks, so I'm gonna work out to figure out how to come over there and shout about it and and hopefully, sometimes whom, to kind of start figuring. Some of this out.

J

As well, I'll actually follow up and make sure that the TOC come with an update.

C

Yeah prepare like in the next caustic for ready and those concerns, because, if I would really like to, if it would be playing the devil's advocate here, I would say you don't have software, you don't have scenes, you have member organizations and they hug for all the endpoints they might integrate. You haven't talked to many of them or some of them don't think we. Yet if this project wasn't coming from the CN CF, we wouldn't even be talking about it. So just keep this in mind when you, and by addressing these concerns, I.

E

Very much do and.

K

I also know my I just want to pitch in for the medium-sized small company, we're very familiar with helm, and we do all most of the things with home and that I think that having one place that combine all the artifacts and be easily the search will help us as a small and very limited resources company to install more objects. So Falco is great over the coop here.

K

It's a great project, but come in and understand how to install it and how to find rules and defines all the artifacts for every project itself is very hard and very time-consuming and the initiative of creating one place, and it will help and possibly take question from like the Falco community. So this place will be great.

C

And I think nobody's debating that centralizing, making things easier discovery for the end user to.

F

C

I just want to make sure that this project gets treated like every other sandbox project out there and we're waiting the same concerns and not saying we're not taking it in there with well as the question it isn't mature enough and has it done enough of its homework or might it take a bit longer but I think Matt you've got my points, we'll have another discussion where you can go and stay around from there. It.

E

May be one of my asks of about sandbox projects is, they are for experiments they're not for mature software and the sandbox documentation talks quite a bit about the different reasons for sandbox projects as far as these kind of experimental things, and so just keep that in mind. As we look at sandbox projects that they're not expected to be like incubation projects, they are expected to be a lot less mature and all of that and there's a whole document that talks just about this.

E

That's been updated recently in the TOC repo, and so that's the only thing that I ask is when we look at this and any sandbox project that we keep it to the sandbox criteria for for those things, because sometimes I know incubation level criteria start to creep in as far as expectation and asks that may just be premature for the projects and that's okay. Under the sandbox level. That's.

A

My two cents.

E

A

Of course, but it's done made it on a meaning that it means that you can be linked experimental project. That I mean that you basics I mean the project is okay, because there is a lot of quite hero there to say: okay, it can be experimental, but it should allow you to sing the ignition and it should be a fun dimension from the fun foundation for successful in datian label project. So I think that'd be the whole statement, so Vicente Fox project.

L

Yeah I meant for me, you know looking at it from a distance, because I don't have a horse in this race at all. Is that I, don't think math that the objections are around at maturity? My sense is that the objections are around potentially around some of the other elements, not maturity, project itself, but are around. Do we have, you know, do have we thought about how this will come together, how it impacts some of these other projects and so on? So that's just my feeling on this.

L

A

Not saying it's a you know, excremental playground for the project, I think maybe some reason understanding their part.

E

Yes, absolutely and I agree with those things, but like for example, I know some of the sandbox projects have been asked to do things like the incubation work, overtime and start getting into the incubation due diligence to go for sandbox and, in this case I'm not even really asked talking about this project, which isn't a requirement and that's a lot of paperwork for somebody to end up doing to go for sandbox, because I know we quite often jump to those so I know we're two minutes over, so I should probably shut up now. Thank you.

E

All for listening and I'll have hopefully a well prepared presentation for you in either. The next meeting were the one afterwards trying to address some of the feedback here. I'm gonna, try and document it on the pull request as well. So it's documented for the future if I miss something on there which I'll be doing after this call, please feel free to jump in and add the points that I missed, so I can attempt to address them. Thank you.

C

And I think this puts us to the end of today's meeting. Rightly.

C