Cloud Native Computing Foundation Technical Oversight Committee, 26 Jul 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2022-07-26 TOC Meeting

Description

2022-07-26 TOC Meeting

A

Thank you all right. We are now recording go ahead.

B

uh Hi today is july. 26Th today is a sandbox review. So let's get started um so we had two projects from uh previously that we had asked attacks to give feedback on, and I believe the first one was hexa um emily. Do you want to um tell us what.

A

um So some of our feedback back to hexo was one. We wanted them to go talk to the security tag so that they were aware and engaged in those processes to engage in a self-assessment which they have already filed an issue for um as well as answering some of our questions specifically around what is hawk. What is the difference between hexa and idql?

A

They did provide that there is a thread in our channel with that information and content. So, overall it appears that they have satisfied all of our asks. I think the only item that would remain outstanding is providing the clarity that they gave back to us personally back into their own project documentation so that any potential contributor or community member interested in leveraging the project actually understands the differences between idql hexa hawk and how all those components work together.

B

Okay, uh any questions for emily.

B

Okay, uh we're ready to vote uh like it gives me the feeling that yes, uh they've stepped up. They did what we asked them to do: they're getting good feedback, they're able to talk to people already. They know some folks in six security, so I think they are well on their journey. So I would you know, I'd be positive on them joining us.

A

Yep, I would add the caveat that we'd like their documentation, to be updated to reflect the better clarity that they provided to us about. The various projects.

B

Yeah we'll have to catch that incubation uh emily, and we can also add a note um that when we do the plus ones when we send out the email plus one and then add some comments there on making sure that they update their documentation. Yes, amy, thank you so amy. Could you please call for a vote on this.

A

Is open, go ahead, yeah.

A

Okay and yes, we can ask for them adding things around now, we'll ask for it in onboarding.

B

Yeah welcome so the second one we want to talk today is uh conveyor, uh I believe, uh erin. uh We got some feedback, and this is the updated language uh that they provided afterwards. Afterwards, some of the key things that I noticed was uh hey. If there are things that are uh open shift related, then we would move it to some somewhere else and we'd like to focus on just pure kubernetes here. um That was definitely one thing that I could catch uh erin. Anything else you you wanted to highlight here.

C

Yeah, I think they took out the open shift specific things because it is based just on kubernetes. I think it was more of a documentation issue and they also broke it down into its other components like forklift and crane and all those things to be separated out. So we have a little more clarity on what is being contributed or not. I think for sandbox is fine incubation. We would probably want to have things more crisp.

C

um I'd like to get a you know, and I know that they went to the tag app delivery um felt like they had good alignment there. So, um and this is the resubmit. What is confusing to me is how we should have them resubmit so that it is an addendum to their original. So we don't have to go back into the archives and figure out the diff. So I did.

B

C

No, I know you did. I just wondered: if there's a better way, we can do that for projects than them go.

B

Correct I have a feeling that we should ask them to email us an update or something like that or open any issue or something to that effect. But yeah. Let's do that offline. For now we can start a vote uh and get this in. I will add it to my uh list of things to think about, uh and you know we can.

B

We can do it ad hoc.

B

uh Amy would please couldn't go.

A

A

B

Okay, so okay, that's two: uh let's start the next one true, so I I hope I'm saying it correctly. True, so um is a kms provider. um There is one more at the bottom that says external secrets operator, so there are two in the same space, but let's just talk about torso here.

B

um I think this was from on that folks. Yeah on that slash also slash wiki.

A

So I want to clarify it's. I don't think it's actually a kms provider. I think it's a plug-in for a kms provider to kubernetes secrets.

D

E

Yeah yeah, um I I I think the scope is like a secret encryption. um It's a scope, very small. I just feel the scope is very small. Does it need to be a separate project and also I do not see the root map. You know, there's a link there, but there's no no map so.

D

There is one a.

B

A

D

Project started in october.

A

Yeah, it's super early um looks like they're, even still working on building out kms provider support. um I think azure is the open pr that they have right now, for it.

D

I mean there's been a lot of projects along the lines of encrypting secrets that have been created over the years that have not sustained development.

D

Which has been somewhat problematic, I think for consumers, which is not, I mean, there's been demand for them, but.

F

And most of the roadmap here is complete, and so even if you go dig through the road map, you're going to see almost everything's closed. So what comes next.

B

Yeah uh here um roadmap for v2, the page is not phone.

F

You you have to go into the projects and go search for it yourself. It's not that the link doesn't work. Okay,.

B

F

If you go to the projects up, there.

B

F

Classic project yeah.

B

A

B

F

There are a lot of things closed.

B

Yeah, so let's just read what they say: what didn't want from us right, yeah.

B

Okay, so the other kms plugins- um probably I don't know so. This is a kms plugin and one of the questions was: was it too small to be a separate project and we are seeing that there's only one person working on it, so we asked them to come back.

B

Is that what the consensus seems to be.

A

I think they should come back. I think that they need to go talk to kubernetes, saying security and maybe potentially tag security as well. um They have very little in the way of community meetup information. So if they're looking to get community contributors, it's going to be very difficult.

B

Yeah yeah, okay, so uh amy uh can we summarize that and send it to them? Yeah.

A

There's a note in chat, let me know if that's good enough.

A

The usual framing that we use for this, like a more robust community.

B

Yeah, okay, all right, we can move on the next one is yam, um so this is like.

D

The github link doesn't work, another github link everywhere it doesn't work, so I don't think we can review.

A

I couldn't review it either. I went looking for it. The closest thing that they have is the yam engine examples and the repo, and that's it it's a hello world. There's nothing else.

B

Yeah, so uh one thing I did figure out was: this: is a person from zoom doing this and do almost like doing it by himself, I presume or herself I couldn't tell so yeah. So let's go back to them and say hey, please fix your links and come back to us when you're ready.

E

And uh another question is whether they can merge into an existing cd project like the other cd or do they need to be an independent one?.

B

Yeah so yam incorrect wavelengths. Yes, thank you.

B

So next one is armada, um so this is from g research.

E

Go ahead. Sorry.

B

And so volcano it's in the same space. I guess.

A

B

A

A batch scheduler.

E

Yeah, it's a batch scheduler and they still so, I think they schedule across clusters and that's a differentiation but their cri and the commands actually are not quite um cloud native. They have their own cri apis, which is their own specific way and also they use radius. They are, I don't know I do it, so I just feel it's not quite, um but they have some very good scheduling. Algorithms,.

E

Do you need this like to be combined with like, uh like declarative model api, or is.

B

We don't have any clients yeah. Okay, so uh looks like you know: there's like a few people working on it for sure.

F

And it's been developed for several years, yeah.

B

F

B

Yeah yeah and the other thing was, um you know.

B

Yeah, I would, I would support this. um You know the technical stuff is okay, they they have a community and they want to do this cloud-native way um and the key difference is um you know it's not a single cluster solution, there's a multi-cluster solution, which is definitely one of the things that we would like people to move towards.

E

Is that uh how many other maintainers belong to one company or and because they're they don't have that email.

B

It's it's from g research in uk, so um yeah and primary reason people come here is to you know, start off to build a community here. So um okay, so any other observations before we can start a vote.

A

um The only other thing that I would call it is they have a fairly significant backlog, but with the project being as altered as it is that's to be expected. Okay,.

B

Okay uh amy call forward. Please it's open go ahead.

A

B

You uh has anyone done a deeper look at cube ray. I I've been talking about things, so I would like to understand.

E

I have taken a look at this looks like it's very preliminary. It has does not have much information. Ray itself is a popular. It's a it's, a popular project, um ai, you know workload, but I think they would like to integrate this with kubernetes, but there's the I think it's very early stage. I feel they need to provide.

C

Race, not real early. In my opinion, it is being actively used in you know many different facets of industry. It was actually kind of instead of spark that was kind of retrofitted to be. You know, used in kubernetes and didn't quite ever work well. Ray was like designed for cloud native use um for aiml, so well,.

F

C

What is it ray? That's being.

F

Contributed is it it's cubrey which lets you run ray applications in kubernetes.

C

Rey is mature. I didn't I wouldn't want to.

E

C

I think you missed.

E

uh I think yeah, I I didn't say, may is it great popular? It's not it's. Not I was talking about. Cupra is very early stage. Okay,.

C

I misunderstood sorry kathy.

B

uh So the any scale is the name of the company. That's backing this project.

E

If it's a berkeley berkeley, I think it's the same um group which is working on ray. It's a berkeley right.

B

D

It's kind of strange taking it out of the ray project and putting it just this bit into cncf when it's already part of a project where it kind of makes it makes sense for it to be. There.

C

I agree, I don't know what the motivation behind that is like.

E

Is it that's very limited, does not provide much information. Yeah ray itself has a lot of information but cooperate. You know not much information.

B

So there is no other projects in the space um is basically what they mentioned here.

B

And they want a closer partnership with kubernetes community, so those are the reasons they are coming here.

B

Like do we have operators today for specific things like, I think, there's one more postgresql thing.

D

Historically, we've mostly said no to operators, because we would have so many separate projects and it's not been clear how to manage that.

D

I mean, I think that it seems to me that they don't need to be in the sandbox to have a closer partnership. They can have a closer partnership by working more closely with the community which they're welcomed to do, but leaving it as part of the rey project just seems to make more sense to me.

E

If you want to integrate this, they can use on a community crd communities operator, something like that. Yeah.

B

E

They are already.

B

C

Yeah, they already are doing that. It's just odd, because I I think it's already being widely used in the kubernetes community, and I don't know if the sandbox component is just to formalize. The community structure.

A

A

I would agree um having it remain with the ray project.

B

Yeah, but if they really want to come here, are we stopping them.

A

Well, it sounds like we have precedence in the past that we don't accept. Operators is that correct.

B

I don't think it's we've written that down anywhere. I.

D

Don't think it's been written down.

B

Yeah see the other thing is like this is essentially the scientific community right like um profs in berkeley, and you know the the general sf uh different uh stanford and other people too right. So we need to draw them in if this is the way to draw them in you know, I won't mind.

D

But we're not, I mean we're not going to get the whole right project, probably.

B

Which would be.

D

Like I would make sense for ray to be in somewhere in this foundation, perhaps yeah, um you know, I think that would that would make sense, but just one little bit of it and not the rest of it is just kind of you know the least the least important part of it is in the next foundation and the rest of it is not it's just weird.

B

Okay, so uh do we want to take a vote, or do we just say not at this time.

F

E

We want to ask them.

F

E

If they can provide more information about what they they're going to, they said they are going to provide tools to do you know all this, but there are there's little information on. You know how they're going to do that. What's their you know, uh so I just feel it's there's not much information to evaluate. I think rate itself is a very good project. uh It's a great project and also I agree with you dims. You know we it's good.

E

If we, you know uh include like those berkeley, you know those universities yeah into the csf, that's a great thing and just if they can provide more information, I just it's hard for me. It's easy. You can find a lot of information about ray, but for cooperate. What exactly you know not much information.

B

Yeah, that's that's the pause part of the reason why they're coming here right like they want to. They want more people to know about cube ray I'm thorn.

F

So I look at this and I I say, being a sandbox project: isn't going to give you all that extra visibility, where everybody's going to notice you.

C

That's what I.

F

C

Just say: there's no marketing for sandbox anyways. So if the motivation is marketing dollars or visibility, that's not going to happen in sandbox either. I totally agree.

F

And what's their motivation, I really want to know why they want this to leave the ray project and join the cncf, and if it's just marketing, I don't think they're going to get that level of marketing they desire. Anyway,.

B

Yeah, that's basically what they've said here right, um but they've said that the.

D

Whole open source, the whole ray project would like to have a close relationship, not this little bit of it.

B

Yeah, so uh I would say, plus one to like figure out: go talk to them in some form or fashion. Maybe they are coming to detroit. um We could have a talk with them or something, but it's probably early for us to look at this. uh The way it is structured right now, so um let's go with no now any objections.

B

Okay, good um next one is open. Zero trust security platform uh which.

D

Is actually no effect.

B

D

The new one and you know have you pronounced it? Yes, although the code, the code is not yet under the org, which is very confusing.

B

A

Yes- and there are some bad github links as well,.

D

They also have not imported the history into the.

B

There are three repositories and you know there is just links to the other new vector, so they haven't cleaned it up. It's not ready for us to review. I think.

A

B

um Matt any um any words here since this is from suse. I think.

F

um You know I'm not sure what their timetable is for bringing over the history. I think they're going to just move the repositories over.

B

They're, seeing that here.

D

The history is not in the in those repairs, yeah.

F

D

F

New repos, it's in the old locations under the new vector org. They.

D

Have not reorganized? No those don't have the history, they just import, they import all the files. As of um last december, ah yeah.

F

Yeah, it doesn't have the full history.

D

Yeah, which is kind of annoying- I mean I I I can see there might be reasons for doing that, but I would write I.

F

Don't have the reasons for it, but I'm happy to go chase them down.

A

I suspect that they were waiting on a sandbox decision before they did anything.

B

Yeah, we need to turn the situation around, so please take that back matt and let us you know come back.

D

B

B

Okay. We are seven thirty minutes in sre works. um This is from alibaba.

D

I couldn't work out what it does from the dollar.

A

Same there, there isn't a lot there. I got the impression that it's for operating and maintaining big data and ai deployments, specifically that are leveraging ai ops, but there's nothing else. That's there. It's not clear whether or not it's kubernetes native and specific, or if it's for a cloud service provider, it's there's just not enough.

B

Yeah and all the documents are uh definitely in a language that we don't read uh well for zero for that matter, and it's from tencent.

E

Oh, it's from alibaba.

E

One concern is their code of conduct, is the enforcement is enforced by alibaba project team? Is that okay.

B

We work through, we have a set of things that they work through as part of onboarding, so we can clean that up. That is not a problem as such, but we need to know what it does and there doesn't seem to be a way for us to figure that out here.

C

It could just be a language barrier, though, on the intent of sre works as well like I like the idea of some more uh projects coming in where it's addressing these operational concerns from a community standpoint, but I just there's not enough there that I can understand the extent of what is being contributed.

B

Okay, so can we redirect them to tag app delivery and ask them to give a presentation there and have a chat and then come back to us? I.

A

Think I want to be able to have them reapply with like robust documentation and project goals.

B

Let them talk to somebody first uh amy, so.

A

Yes, and one of the things that we've kind of committed to projects is when we do that, we have specific goals in mind for the tags to be able to review, and here I'm just seeing larger clarity. Yeah.

B

Okay, how do you want to phrase that amy.

A

um I I really think it's like the we're unclear on the project's goals, um we'd like to have a presentation around the goals, the road map and how you'd like to get more contribution.

B

uh Maybe it is there, we just don't know how to read it right, uh since we don't know, okay.

A

Look I like richie's she's rooting for this offer an intro and a demo to the tag perfect, yeah all right. Thank you. Okay. Thank you.

E

And that was awesome to provide this con all this in in english right. They need to yeah fix that.

B

I won't force them just for us to review, but at least let them talk about the technology and what they want to do and get a summary from the tag.

B

Okay, thank you bumblebee. Did anybody catch how this works and why solo is bringing it here.

D

So it seems to take ebof ebpf programs as oci images and load them into your kernel. um I'm I didn't get.

A

E

So I um so I think I I reached close. I went through this project. It's it's targeted as build and run this ebpf and start containers um say quickly. um I think this ebpf is a very important.

E

You know trend in the a way of doing networking, um so I think this is very useful um yeah, but in terms of information, um not very much information, but I think this part itself is.

F

Right, I think the goal was to make ebpf have a easier user experience for people to develop against, and that was the goal of the project.

F

I was wondering whether this is something that fit into the cncf being, where ebpf sits all around right. Now they only support c code for what it wraps, although they do have rust plan development only started last november, and since february there have been 12 total commits since the start of february, with less than a thousand change lines.

B

Yeah, the tutorial is really nice. You get a feel of like what they are trying to do here. um How do you write a new ebpf program for sure, but clone cloud native connections are mostly around like hey. How can we package this into docker uh as a docker image, oc image, kind of thing right.

B

uh So going back to like oh, what are they.

A

Is there any reason we can't refer or recommend them back to the ebpf foundation.

E

I think they are not specific about ebpf, but they are more about you know, tools or how to build and run it run the ebpf as program as containers, and I think that's why I probably want to come here.

B

D

Yeah, but I think people are allowed to package things as containers outside outside cncf.

D

I don't think I don't think that necessarily means that they belong here I mean I would want to see more community adoption just because it seems that right. um It's not clear that that you know that is going to be the way people want to package things so.

F

And it's ebpf and c code, which are two things the cncf is not really known for so I'm I'm I'm looking at this from it. Does this fit with us.

C

Well, what would be similar to something we've had I mean we've had recently some plug-ins to like vs code for packaging, I'm just trying to think of are there?

C

Is there tooling that's similar to help transition things to be cloud native or run within this ecosystem that we have taken before.

B

Closest thing I can think of is like code generators for things.

B

So that they can be deployed with uh kubernetes or cloud native stuff,.

B

C

Do we have any update from the observability tag? It says they didn't present.

C

What I want to avoid is feedback we've been getting from various community members around. They go to the tag they present, they get really good feedback and we they come into this process and then they get bounced and they feel kind of like. Then they have to reapply. It seems like a frustrating vicious circle, so agree.

B

I agree so we we can give this a home and see where it goes right.

A

So my question, then, would be why bumblebee at home here in cuba, not.

C

I'm not advocating one way or the other I'm just trying to figure out just noticing in the comments that they had already presented in. If.

A

I completely agreed, I think, part of what we're starting to run into at least based off of this sandbox discussion is: there are extracurricular projects that are being proposed that aren't tightly coupled within the ecosystem, they're more uh natural extensions of its maturity. So the question is whether or not those extensions which kind of circulate in a different project arena whether or not we should be including them or if they should go to that those other project spaces.

A

And I'm not sure that we're ready to have that conversation to make a definitive decision, but we should probably do that soon because it will impact future sandbox applications as well as the existing projects we have in the ecosystem.

B

So let me play one argument right. What we do have projects that do awesome and then wassum gets uh deployed as containers into various environments, and we do have a bunch of projects that doesn't do that. So this is replace awesome with ebpf and essentially.

D

But I don't show that ebpf has the same characteristics as well as um that make it interesting it's about a cloud native ecosystem like there's a single shared kernel, yeah.

B

D

You're putting injecting code into and you can't run general applications, it's a it's a tool for building.

B

D

Tool for building kernel instrumentation and for building monitoring tooling, and things like that, which is not in that sense cloud native at all.

B

Okay, so that's what we tell them: we we tell them that. Yes, we have seen things similar to this before it's just that this is a code generator for building tools that can be deployed uh that you know somebody runs this code generator once write some tools and then release us in the while, and uh you know off of the girl, but that's not workloads as such right. So that is the distinction that we are making here.

D

Yeah, I think that using cloud native packaging doesn't make what you're packaging cloud native.

B

Yeah got it okay. uh Can you write this up just yeah.

A

D

With that, one.

A

B

Thank you. um So next one is cloud native pg uh who is passionate about databases in kubernetes.

F

Well, I mean I'll I'll start with this one. So cloud native pg is a postgres operator for kubernetes um and that's essentially what it is, and if you look at it, they do talk about the other ones out there crunchy data things like that. The different operators out there and they're looking to have one. If you keep going to the right, you'll, see uh their motivation here and they're looking to instead of having it disconnected having it connected. But at the center of it is it's a postgres operator for kubernetes.

D

So are they asking us to king make it effectively by promoting it over the other projects, not that all the other projects would like to merge into this one, but they would like us to king make this one by putting it into cncf.

E

Yeah so yeah, so kubernetes has quite some operators right, crb and operator. So I um I don't know why. Why do they want this to be a sense of suppression, project.

A

This falls into the same discussion we just had about operators.

B

I think here what I'm seeing here is like. There are multiple postgres vendors. They want to work together on a common operator um here. Well.

D

I I don't think that's what they're saying I think what they're saying is. They would like us to choose this one and force everyone to work on this one. I don't. I don't see all the other projects saying we want to. We want this one to be the one that's chosen.

E

D

So like, if all the projects were saying, we want to come together and work together on a project and we need it in a neutral place to do so. But the way I read this is this project is asking us to be a king maker and choose their postgres project over all the other ones and not let the other ones in because then everyone can work on this one which is not what we do.

F

And almost every developer is from a company call from one particular company.

B

F

They're, all from the same company.

B

Yeah enterprise db okay, so um we should tell them to go talk to other post-rescue vendors, get consensus and then propose a clean slate or or something.

D

Well, I'm not sure we would accept it. Even then, okay, given the other discussions, so.

A

If what they're purporting is to kind of bring everybody together to do this level of development, would it be beneficial for them to seek a working group within one of the sigs or one of the tags, to start having these discussions? Rather than doing this key making request.

B

So what do we tell them to go? Do tag up delivery.

A

I I think personally, I feel like, if the, if that's what they're they're seeking based off of what it is that they are saying then yeah tag, app delivery, propose a working group, get the right people involved and see what happens. Okay,.

B

uh That's fair, so um can you write that up emily? Yes, richie um is more than an operator. We should. We wouldn't be stopped from accepting a different project. Yes, you are right. Richie.

D

No, we wouldn't be a stop, I think, but but the rail reader is they're asking us not clothes are in effect, which is.

B

Yeah, so they have to adjust their expectation that we'll we might be bringing in more right if we do decide to do it. That is.

B

B

So we give them feedback and move to the next one.

A

Actually skip two down this.

B

A

B

A

B

uh Db pack, I think I had trouble with this one too.

B

So this one uh they're talking about using postgres, uh no mysql protocol or something like that.

D

This project was started in april. I think we're getting to a record of how quickly it's been from creation to submission.

B

Yeah, so I think they want to support multiple databases but doing what db pack a database cluster tool pack tool.

B

Yeah this was the one I was talking about. Mysql protocol.

C

Does it create like a layer to make that consistent? Amongst all databases, I don't really.

B

C

I don't again understand, like the full extent of it,.

A

Says that it creates an operator to manage database topology, manage traffic and allow users to customize traffic routing through hints using crd rules and detecting hint and sql.

B

To face commit.

A

They want to prevent other companies from launching their own database mesh projects as well as get more developers. That is what their request is.

B

Okay, so that's not a good way to talk.

B

That might be language barrier too right. uh It's a transfer.

A

B

A

The only the only other call out that they have in there is to form a standard in the field of database traffic governance, but we don't I'm pretty sure we don't do standards here. We need.

B

Yeah yeah yeah again, that might be like a.

D

I mean it's just too it's too it's too young that it literally is a month. Okay, oh this like so, we can actually just tell them to come back.

B

Yeah we can tell them yeah gather some more momentum and then we can revisit this.

B

Okay karina, csi, plugin, okay, who's, the storage person here.

C

Again, this is like our in more the database scenario, which is interesting that that's kind of heating up.

B

Beyondcent.Com.

B

Extremely low latency battle tested, building blocks, well-known dba experience.

C

Again, it's I'm the goals of just marketing and bringing more people in. Don't aren't quite what we're looking for.

C

And I've never heard of the cloud native storage system that they're contributing here. If that's paired with it or not,.

B

Which one are you talking about.

C

Let me find it scroll to the right. I don't and I also don't think we have each of the csi plugins as separate projects. Yes, it's kind of just a subset of the csi specification for cube, so.

C

I I guess off the cuff, I don't feel like it's something we would generally take in. I I am not sure that they have presented to the storage sig either.

B

Okay, so then it's easy go, go talk to the storage sig.

C

Yeah and and have you know, alex provide us a recommendation from that, but typically in the history we haven't taken csi plug-ins as a separate project in and of itself, and maybe I'm missing something that that's.

D

It's also a kind of new recent project without much history.

B

Okay, so easy thing to say there uh go talk to the sake, come back tag and come back.

B

C

But again, if we say that I would like to in a separate meeting figure out how we get the feedback from the tag. If we've made that request.

B

C

Need to be able to populate this spreadsheet with a reapplication and a recommendation or whatever from each one of the tags as part of our standard process. Yeah.

B

Let's um do an async thread on slack uh after this meeting and we can decide there. um Hopefully you want to get the feedback.

F

And we have to decide: do we want csi plugins to be individual projects in the cncf.

E

F

And if the answer is no and they go present to the tag and the tag says looks good and they come back to us and we're not interested in csi plugins, then we just ask them to do a bunch of legwork for nothing.

B

uh So we have to tell them what we are looking for. uh We have to tell the tag what we're looking for right so then, but more.

E

Agree with matt, you know the. If you know, if we we don't think you know csf, plugin or any other plugin like you know: ci plugin. Okay, there are many ci plugin that many csr plugins. If we don't think that should be separate project and do we still need to ask them to go to the tag and then come back yeah or if we allow that there will be a lot of plug-in. You know we're gonna see but yeah.

B

We have seven minutes, let's leave this for now um as uh something that they need to do. uh Let them go talk to the the tag and we'll also tell the tag that is it just a csi plugin, because we don't want just a csi plugin. If there is more than like, uh there might be another language barrier here, they are highlighting one of the aspects of the thing which is a csi plug-in and they might have other things like.

B

I saw a scheduler or something too, so it might be like a full solution out of which they are highlighting the cloud native aspect so um giving them the benefit of the doubt here. um So multi-cloud uh anybody saw this one.

A

It's basically terraform once deployed everywhere.

A

D

Yeah I mean it's a I'd say I mean they started in january. It's a it's a difficult problem like it's a doing. Multi-Cloud implementation of terraform is like a huge piece of work.

A

Their commits have been pretty active, though.

D

But it's two people in six months I mean, I would still say, come back.

B

You're talking with the team here, um looks like they're five people- oh they have. The name of the company is the same as the name of the project too.

D

There's only two of them with more than two commands.

B

Okay, yeah you're right they should. They should grow a little bit more community-wise before they come back here yep. It seems an interesting idea, like you know, generate terraform and then use the use.

D

Yeah, but it's it's, it's not terrific. It's not! It's not a trivial.

B

Problem like no yeah, it's a gigantic.

D

Problem and so it's a question of like- can they create a community that's up for solving a hard, a hard problem with lots of difficult edge cases and.

B

D

And and and that's the kind of that's the question really rather than just like yes, it's crazy, it.

B

Works right, like you, will need to drop down into a terraform for any edge case, and you know soon. It'll be useless kind of thing right, it's difficult too not every resource will be supported.

D

Or your support lowe's common denominator, so it won't be useful. I mean it's like there's a there's all these issues that uh they'll have to have a opinion on okay for, for a start like just what are we? How are we going to try and do this? Given it's a hard problem.

A

I would say, in addition to reapplying with a more robust community that they should pick a few particular use cases that they intend on doing really. Well, because that that's the only way that they're going to drive some of that community adoption and development.

B

Yeah, okay, um who wants to write this one because we are saying no but come back, but also adding the context that the two of you just said: okay, uh amy sure! Thank you!

B

uh I'm surprised! We got to 14 there. You go uh four minutes left and external secrets operator. uh Did anybody get it so? The main thing about here is they already brought a whole bunch of pro people into this project uh and they are very proud of the fact that uh they've so, for example, here, uh if you're curious about the origins of the project check out this issue on pr, so apparently there were a lot more projects and they all gotten consolidated here.

B

So these standardized crd spec and uh you know, there's a links to a bunch of other operators here, uh so they basically bootstrapped already into a community, and I think it's a good good place to start. What do you think emily.

A

They're also an operator.

B

Yes, they are also an operator.

A

I mean, I think we should be consistent.

B

Yeah but secrets are a big problem.

D

Yeah yeah, I'm not sure we should be consistent with this one, because it's, um although it's implemented as an operator, it's not just an operator to exit a way to manage secrets in kubernetes, which is a which happens to be implemented as an operator.

A

Yeah, it's very I mean.

D

And it supports multiple.

B

Yeah, it's not one thing that they support right, yeah,.

A

The secret space needs a lot of work for us.

B

A

Like even all of the things that they have listed, not a single one of them is a cloud native project.

B

What do you mean? The providers here, yeah.

A

The providers, aws secrets manager, hashicorp vault, google secrets azure keyboard, like.

B

What have you won that openstack? I.

A

Mean I I would love for us to have our own secrets management project, but I know that that is probably not going to happen anytime in the next few years.

D

We have, in the past, talked about how we should how to try and get one.

A

Maybe maybe this is how we do that.

B

Okay, uh people who are watching this video, please start one and we you will have an easy path into the sandbox.

E

We had another secret uh operator right secret operator before right, so yeah.

B

That were just two people or one person yeah. There was only one person working on it. uh Oh yeah kathy, so.

E

If we want to have such a parting csf, can we ask them or to consolidate.

B

E

They're doing here like.

B

There's a I showed you that um okay.

D

Yeah that list they've got yeah is what I raised on the other one. There are all these dead projects and activish projects. I think, but this list is kind of part of the scope of the issue that people don't seem to be able to get momentum behind these projects long term for reasons that are.

B

Right but this one is getting momentum right, yeah, looking at yeah.

D

At least for now, so yeah so I'll be inclined.

A

I'm okay to try. It.

D

All right, yeah, okay,.

F

Yeah I mean it has a community, um it fits the sandbox which is where they're trying to work it out, whether or not it will last or not is on them, um and it plays well with a lot of different systems which I I know this is a whole people want filled so there's potential there I mean it makes sense.

B

Yeah um so amy, can you start it's open. Go ahead! The last one for today.

B

Thanks everyone, we got through a whole bunch today and hopefully quite a few people will be happy.

A

Gonna track towards the next one being september, 27th any objections.

B

A

B

Okay, thanks a lot everyone. Thank you.

A

B

A

Okay, thank you.