Project Harbor Community, 1 Jun 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CNCF Harbor's Community Zoom Meeting - June 1, 2022 - Happy Kid's day!

Description

CNCF Harbor's Community Zoom Meeting

A

Okay, hello, everyone today june, the first, uh it's the international kids day, so happy holiday, everyone, um it's the community meeting for project harbor, uh it's official cnc meeting. So please follow the code of conduct for sincere.

A

What I'm sharing right now is the agenda for today, it's quite clean so far and please I've shared the into the chat the link to this file. Please add yourself if you like it and add your topics to to the discussion topics list with that. um I know vadim added few of this. So do you want to take away.

B

Yeah, um first off um I saw that this this week there was the released two two five one, and I already saw the nice thing there that's immediately spring into my eyes, or is that was uh pleasant to see at other release notes, and I have to say that the release notes that we see in I mean there was already 2.5 rc1, but in 2.5 we can see the release notes. Let me uh yes yeah you can you can click here exactly? This is beautiful.

B

This is really a progress, so it is like without much effort. We can see okay, what to change uh what did improve since uh the last release, and it's really wonderful. So the only thing that I would like to change is this component updates, maybe the section here, the second one which is called component updates and it actually contains everything that enhancements.

B

So I would no. I think that it does not. It's wait, wait, wait. It says so. The tickets in there have the tag and the component updates have yeah an update yeah. So I would maybe yeah write updates, but it's just a minor minor change here um other than that, it's uh it's wonderful. So it's really low effort to get the release notes. I like it very much.

A

Yep, thank you, everyone for supporting this effort, and um you mean we should tweak a little bit the labels for.

B

For this stuff, well just the title: it has let's call component updates, but I think it's not clear what this means component update, so maybe just uh call it updates or changes or well or improvement or yeah, because okay and enhancements is, I would maybe call enhancement like new features, because it's also alliance with the label, so the label is called.

B

um The label is called.

A

B

Enhancement: okay, that number should keep it similar, as so, let's keep the enhancement and the updates yeah. But it's just minor details. It's nothing.

A

Yeah these sections and the corresponding um labels were initially proposed, so we have our first run and if everyone like agrees that we can tweak it a bit, I'm super happy to to address that um so yeah, that's that's! Actually the first try. I think that we we do this one. So.

C

A

Right yeah, I think that it.

D

Yeah that release knows it looks much better right in the thanks for make this happen. Yeah.

B

Great okay, um so I think the next one is also on my agenda is or on our agenda uh is the recap from the cube card, so all in and and I we have been at the kubecon in valencia.

B

uh I think it's already one not two weeks ago, right, yeah.

A

B

Yeah so yeah, um it was really nice a lot of people. um A lot of people were coming by at the booth of harbor. I know I yeah a lot of discussions from people a lot. People just came by and say they're just really high hi we're using hubba, really like it's cool product, um and there have been a few people as well. You know that that stopped by and and started discussion with us and um so they've been quite off of big user base of harbor.

B

So a lot of big companies who are using harbors harbor, and there are a few uh big companies in mostly in europe as well, because you know most of the people that were there are away from europe and like governments from, I think, belgium, netherlands, uh I think in germany there is also an effort to build a harbor container registry for some governmental institutions.

B

So there are quite a bit quite a few big companies. You know adopting and using harbor and um from the discussion I would say that that took stuck into my mind was on one hand the topic topic regarding the edge use case. I think this. This was quite there have been quite a few top people interested in this. In this edge use case of harbor, um there have been a few people that are we're interested in integrating or using harbor with other tools.

B

You know with the cicd tools, with the signing signing harbor so like zig store, for example, and some similar solutions. So this is something that I had a few times the discussion.

B

um Of course, there have been a few questions regarding uh some problems with the harbour, and so the other question that I've had also a few times was basically how to make harbor work in multi environments. So this means like, if you haven't, want to deploy to death stage prod. How can you separate this process, or how can you map this into in in the registry, and we had a few discussions with a few people around this topic, and so one one proposal was to have multiple harbor instances.

B

You know one for deaf one for infants for for, for, for prot uh other approaches, you know with the zig store, you know with the signatures. Signature validation on the on the cluster side, so there are quite a few discussions.

B

um There have been a few people using harbor in combination with good helps, and there was a few problems there as well in the githubs context, but I honestly cannot recall what type of problems people had there.

B

So, overall it was, you know. Quite we had a quite positive feedback or received quite positive feedback from from the fox who popped by, and you know, said, hi yeah. So do we have something to add all in.

A

Yeah my biggest takeaway was folks were I'm not sure if you mentioned that I was uh talking another issue. uh The I in my conversations folks were like yeah, it's great, but can we have like a normal working apis? So this is yeah.

A

So that's just like the biggest issue for folks that they want to automate so many things and right now, uh due to the api, it's not fully functional, they they have issues and for some for some of them, even that was like a show stopper to adopt harbor completely and they had to switch to some other registries like key or yeah, or even some public uh like a github or whatever.

A

So I think yeah that was like every every second user that was chatting with me and he was like yeah apis. So.

B

um Yeah, so there is that's, I can confirm this. I had also a few discussion regarding api and then the following topic is the permission sets like how can we uh do if the? If so the companies are starting using the apis and then the next question is: how can we restrict access to certain things? How can we do things with the api, for example, with the robot account? So there's like a lot of questions that can we do this with uh with the api?

B

Is it possible to you know, do things so this was actually the the follow-up question after the api like how? How can we express uh certain workflows with the over the api, um yeah, so yeah, so something like project creation, project management, uh the pipeline uh pipeline things as well: um yeah, okay, yeah, okay,.

C

D

Yeah, I think yeah that's very interesting and a lot of feedback, so I heard uh vadim you have mentioned. There is some some I mean some people in the even to talk about the harvard page right. So is there any special use cases they describe that they need the harvard age side.

B

Yeah, so I'm currently trying to set up a document where I'm I have a few different ideas. How does this? How the other requirements on the edge side- and I wanted actually to reach out to you guys and see- um is there already something on in works of uh of harbor from the past that you have that you could share with me.

B

So we could maybe uh take a look at this together and I can build upon this because otherwise I will start with a fresh, fresh paper and I had a few discussions on cubecorn and I have now a few follow-up discussions with a few uh people who are who need this use case. So basically, there are different use cases. You know and they're trying to identify a common pattern that we can start with and then allow us to expand in different areas.

B

But I'm trying to understand what is like the minimal requirement to have this edge use case working.

D

Yeah, I think we we originally have planned some investigation research on in 2006 right to finance to identify the requirements to support the age cases. So if there any uh use case can could come in from the community, I think it will be helpful to I mean for for the hub project, to I mean to identify uh what is the common use cases to support yeah? Okay, so anything we can. We can follow up this and offline to get something.

D

uh I mean write down to to discuss right.

B

Okay, yeah I'll start a document, so we can kind of contribute together there a shared document, and um I will also invite a few other people who have met at the kubecon to participate in that.

A

um Yeah yeah, by the way I I think that topic circulates a bit already and I think alex shu when he was uh at the role of pm. He was saying that we have something and uh can you get in touch with roger, maybe maybe he he can.

D

Yeah yeah rosie. I also tried to uh collect, uh collect some uh requirements for age from vmsaray. We may also have some use cases to for hubble to running an in-ear cell, so we just want to see what what is it other use cases to for age users for harvard to use to support edges error from community, so we can to identify what is the common use cases that the community needed.

A

Okay. Well, but as far as I can recall, alex was sharing that we have something like in, as with him mentioned something in a working like, uh like almost ready proposal to be to be submitted.

D

Yeah, I think that maybe a bit of difference um set it up back to years. I think, um but anyway we can, we can see yeah. We can connect this again.

A

Yeah and I have a chat tomorrow with uh with roger- I think it's tomorrow, so I'll- try to uh to bring that topic up and and try some some info out. Okay,.

E

And and invariant um alex here, um if you need anything also from from my side like requirements or something else, just let me know, and we can.

A

E

It in exchange, because here it's mercedes-benz, we also have a lot of customers which which asking about the feature- and um I think we have also a lot of requirements and api calls which we would like to see or which our customers would like to have.

B

Okay, yeah I'll I'll loop, you in then okay great. Thank you.

A

Okay, go ahead.

B

Okay, um my next topic is uh regarding labels and the question: how many labels can we have? And the thing is I already um I already started deleting a few labels. Mostly, I deleted the labels that are um that don't have nothing so that labels that don't have an issue attached or something sort of completely unused labels. I've started deleting a few of them because if you take a look at the labels, let me let me show this um yeah.

B

Actually, I'm I'm showing the listed labels, which is oh.

A

B

But you can expand this list, you know. So if you, if you, I think, there's like the extra there's, an extra yeah labels, editors, yeah, exactly 278 labels.

A

B

Yeah and- and I would and it's kind of really confusing to have so many labels, you know and now that we have these different areas, so I think the areas. So I think we should delete all labels that don't have any uh any.

A

Issues or or yeah- and I I think we can vote this right now- uh to remove like all the labels uh that are not attached to anything yeah. Okay, so can we do a like a formal voting over here.

A

If you are pro deleting all labels, please raise your hand into the zoom.

A

Can we do that? I think we can or something right in the chat.

D

Like yes, you can, can we filter out uh the label without any issue attached so.

B

D

B

Like, for example, app ops, you know like this label is, this could be deleted or every access locks. You know.

A

That column in in the middle shows the issues the number of issues or or prs which the label is attached to. So, if you sort by, for example, was that most issues.

D

Is that issue lambo, I mean uh include both close the issue and open issue.

B

No, it I think it's just. It just uh contains issues that are not attached to anything, because the closed issues are there yeah yeah. If, if you have an issue which has a label, it's it's.

D

It's there so yeah can we click a label to to see. uh I mean filter the issue with with this uh kind of label to say what it really looks like. ah Can you repeat, I didn't get it. I mean. I mean if you filter the issue with a label it will list. All the issue include the closed and the open issue uh have this label right? Yes, so can we use the label as an example, so you.

A

D

This one is that.

A

D

I mean not the label with the issue I mean, have a label, I mean it looks like a main label without the issue attached, but just to make sure that that is really no issue attached to that label. Okay,.

A

You can you can select one yeah.

D

A

Okay, so let's see sort by fuse and then just pick one yeah like.

D

Like absorbs, I, I don't, I don't know whatever yeah anyway. There is one closed, however yeah you will see. Actually there are usual attachments that because it is closed, so you're not sure the number right.

B

But do you need to close the issues? Do you need to close labels on the issues, because you know that you, you see the uh you see this yeah.

D

B

Yeah we needed to assist yeah. I think we needed it because we.

D

We have some issue with the label yeah.

F

Should carefully to remove some labels and from a closed issue is someone is searching closely to find out the some solutions that we provided in.

A

F

The maybe in the past several years- and I agree that uh it is the theoretical uh problem for labels, but we we should carefully remove um any of them, because we, uh I believe any of this table.

F

They is attaching to some closer issues and they are available for users to fill their um closed issues from the course.

B

But what what's what's the value.

F

Of those labels, you know.

B

If they're sitting on the closed I mean I can understand that labels on on current issues make sense, but.

D

I think again just to make a point. uh uh It is useful for search the uh some historical issue with a specific label to find out. uh If, if there are any uh issue we have, I mean have been fixed for some special problem right. So if you use the label to search the issue, it will list out all the issues related to this label, for example, if this label is for just like example, if if this level is area replication right, we search this label, it will list all the issues, including.

A

D

Issue and the closed issue, so we can easily to find out some historic issue was being fixed with some reason or some similar problem right.

D

D

Yeah but uh yeah uh we should just be careful to.

B

uh In future, I think yeah because, like even if you want to search something by label, I think it. You will search quite long in those 278 labels to find the label that you really need. You know like.

F

Yeah, I agree they just labeled two after each arbitrary. So but uh if you we, if we want to remove some of them, so we have to uh do it one by one and maybe we can remove some of the and reverse some of them. um But we should calculate to do that. Yeah, yeah.

D

Also also, there are a lot of labels related to a specific release, right, a target and with a reduced number. So all the issue will be closed, so it will be showing at zero number. If you, you will view the label in this page, but that is useful right. We will know uh in which release we have include. uh What can we issue fixed and uh what? What what the epic and what feature we've done right.

B

um Yeah, okay, in in in the future, you will see it in the release, notes right so in the.

D

What yeah! But that is not easy for search to understand.

B

Okay, I understand.

B

This for now or we can, we can leave it then for now, if there is.

B

The next question- or the next topic um is also for my site, but maybe we can yeah, we can do the pull request and then we can do the same thing.

B

um So if, if we take a look at the current harbor project and then the pull requests, so there is, I think, like 45, pull, requests open and so yeah 45 now, and there are a lot of pull requests there yeah the stale, but they would make sense to to get emerged.

B

And my question is: how can we improve our merge, request, approval rate or increase it so that we can approve more pull requests or what would help to increase the pull requests so that they're not becoming stale and people get uh that that provide pull requests, get their pull request, accepted or denied fast, and so what would uh uh yeah? What would help to um to get this uh to get this working? Because this is like this?

B

This tail pull requests, are, you know quite yeah a few years old already, but there are still a few poor uh oppo requests that are open since the month that are continuously unstailed.

B

So my question is: how can we improve this this situation here, that we either accept or deny pull requests? So what would help to improve the rate, so maybe a test environment or something I don't know something that that would improve the process. What do you think?

B

What could it be.

A

Can I share my thoughts on this one sure uh now uh for the last month. I think we when, when someone creates a new pull request, the maintenance list is a group is assigned as a reviewer, and that was my first effort to actually accomplish what you just described to them uh to bring the attention of all maintainers to the uh to the pr's.

A

So I think if that doesn't work, so I'm not sure and I'm gonna ask everyone: are you receiving those notifications properly and do you pay attention on them and if you don't why and how we can fix that, uh because the idea of code owners and automatic assignments of reviewers is exactly this one.

A

When someone files a pr you folks, like 11 or 12 people to get notified- and you can practically review that in some like timely manner, and I can see a trend that that doesn't happen. So, do you reiterate vadim's question how we can from maintainers point of view how we can uh improve that? Do you need someone to explicitly?

A

um Are you as a as a reviewer or what or to do like a scheduled um like a pr review meeting every every every week or something.

A

It's an open question for everyone, uh because I can give example for uh the community repo and the website ripple that works and when, when uh we get a new pr me or abigail, we get notified so actually both of us and we just review that so for for but yeah, maybe the volume of prc is not that big there um so is there is.

A

Is there a way how I, as a community manager, can help out with this one to facilitate, for example, once a week, a pr review meeting, or can we do it like during our regular meetings, every two weeks, which is a long period? But uh still it's a it's a good start.

D

Yeah, I think the the team I mean. I have a regular meeting weekly meeting to review the newly opened pull request in the past week. So uh team will review those pr's and assign some maintenance to follow up that, so that is kind of the uh process that we I mean. We will folks.

D

Fall out right, so, of course there are many all requests here.

D

I think maybe most of them uh is due to the tr creator. Do not respond, so that should be some uh some uh some pr.

D

Is this case I mean.

A

You mean there is a pr someone from the maintainers team, let's not focus on the vmware folks, but in general for all maintainers in the team uh reviews that and then the creator is not responding to the comments. Is that what you're saying or I I didn't get it.

D

Yeah two things: one sentence is a vmware maintenance folks have regular meeting weekly meeting to review the newly opened pr in the past week right and we'll, and also we'll send uh some owner to review that or follow up that. So that is the first. I mean to see the process that we we are running and the second one is, I mean I mean there's yes, there are many really old poor requests, not a get close.

D

I think some of them, maybe is due to the pr creator, uh do not respond so, for example, if they review ask for some changes or leave some comments, the pr creator did not get that addressed so that api is left here.

A

All right two things here from community management standpoint: uh those meetings that you're mentioning uh this should be open and I think if we can make it during this meeting or we can set up another one, um so everyone from the community can participate on I'm by community.

A

I mean in this case the maintenance team, not only vmware folks, but to include everyone that it's in the maintainers and second thing is: if the one of the maintainers is assigned as a reviewer as a and is a as an owner from our side, he or she must be responsible to pursue the the creator first to give feedback on the on the commands. So, for example, I opened just a random one in this case, I think in some in some time frame, but him should respond back hey.

A

Can you update this and that I've commented something or we can do that some other way to like to have some rotation from from the team um to to ask for for for commands back, I mean it's ours job to um to ask the the pr creators to respond. We we cannot wait on their will. Just okay, I've created something: I'm gonna respond back.

A

We should encourage them by actively pursuing them to um to respond back, especially if, if he's a good one and and we want to to be to be merged, but there's minor fixes now and there, what do you think just throwing ideas uh out.

F

uh I'm not sure whether there is a faster practice for pr reviews in the whole cncf community, but for me, 45 is not a big number for a graduate project. uh If you see the promises, there are more than 200 prs open years, and some of them are open for five more than five of six years and um would be to uh us reveal as a maintenance.

F

We will do the interaction and add our comments to the prs from the community and but for me, if I see some peers, um maybe you should not uh uh reasonable. I just let the pr open- and maybe someone from the community see that pr and they can add some comments, um yeah um so um yeah. That's my.

B

F

B

A

So you say: if you think that the pr is not good enough, we should not comment back or I didn't get it.

F

uh For me, I I will not close that one, because that that was uh uh opened by the other and if I close that vr, which I think is not good enough for us, but I think it's not a best practice.

F

I just let the pr open and then maybe someone from community can add their opinion to that pr. And but for me the pr is not good for us and it's it's not maybe something it's not reasonable. It's not valuable, and but uh I will not.

B

Close that one, but I think it.

F

B

The pr is not good, um however. We should comment this and say: look this. This pr does not uh is not sufficient enough to improving the quality or this vr is. There should be a reason why it is not accepted, because maybe it's of the strategy or it's uh I mean for whatever reason you know, but it should be transparently uh communicated back to the contributor, because you know the person you know spent.

B

You know days and weeks, maybe to create this pr, so they they have expecting something in the sense that their contribution is valued and accepted into harbor, and I think we should communicate that and see like I mean why we are not accepting prs. I think there needs to be a reason why we're not accepting a pr, because it's not does not fit the strategy of the project you know, so it does not fit the overall strategy and goal of the project.

B

So if someone wants to, I don't, I don't know, store npm packages in there. Maybe it's not the right thing to do in harbor and hubbard only focuses on oci images and not on other package formats. You know artifact formats, so we should clearly communicate this in such a pr. But if it's an improvement, if it's a bug fix of something, I think there is no reason not to accept the pr. There is a reason to to communicate to the user and say: look: it does not qualify because the quality isn't good or whatever.

B

But if someone is improving the harbor, then we should you know not ignoring it, or at least ex at least you know, give feedback to the user and say what he has to do in order to get accepted.

A

More on this one, I cannot agree more on what you're.

F

Saying uh yeah, I agree with you. um um If we see the open prs one way, one some of them are already approved by the maintainers, but they are no further action uh from the others. They do not replace their code. Sigma delays is called to make the ci pass things like that and for uh we do uh uh uh accept pr from a community and we do they review the pr forum community from a community. I mean from the montana group, but I'm not.

F

I don't know clear about the pr strategy that we are talking about. Are we talking to uh down to some numbers? I mean for the open prs or uh we are want to have a maybe a cycle for a prs like if one pr is open for more than half a year, we should close it and we should discuss some interaction with houses on.

F

So what do we want to improve on the pr strategy.

B

So I think one one idea to of improvement would be to um to provide feedback to the to the contributors. You know so that there's a lot of we have a new lot of new contributors there who are doing the first.

A

B

And if we don't respond, it's gonna be the last pr. You know and I think from our.

A

B

The community, we should, uh you know, embrace them and welcome them so that it's not going to be their only pr so that they're going to provide more prs and use the product so because most of the companies or people who are using harbor, they find issues and problems there and then eventually it becomes so big that they provide a pr for that and we should not stop them from doing so and should accept it.

B

So my my goal is that we- and if I look at the pr list that we have there are, I would say, 80 of the prs there are bug fixes you know and from my perspective there is no reason to uh uh to not accept them. You know if, of course, if they qualify and and and passed the criteria but yeah, so there is from my perspective, there is a lot of prs that are actually bug fixes and we are not accepting them.

A

F

Just to clarify something.

A

uh The criteria of accepting pr is kind of vague right because it's not well defined, what's a good pr, how the pr should look like uh and the quality of the code and the quality of the contribution, it's a bit fake and not strictly defined right. So if we find an issue understanding that concept, I think that's the, why we're community and why we need a few reviewers right.

A

So if someone is not sure if that's good or not just uh think some of the other maintainers and ask for help to to double check that and yeah is what dim said when I contribute to other projects, uh not harbor, but like the ones that are not I'm not um attached in in in. In this way.

A

That means I've had an issue, and I spent some hours to to fix that thing on my site, and then I decided to contribute it, and I expect someone to at least say: yeah, that's not good, or we fixed it or something to someone to say something so to to answer the question. What is the strategy? The strategy is to be more welcoming in terms of when someone opens a pr to be good citizens and comment and to help out that that pr to be merged.

A

uh If, of course, it's again applying all the rules for uh strategy for the project and if it's not fixed in some or plan to fix in the future release or something like that,.

A

Sorry, you wanted to say something, but I I think I interrupted you very badly. Sorry for that.

F

uh No, I just uh went through all the pr's on we respond, almost all of them and for the prs to fix bugs. I think some of them are not bugs and- and it's not reasonable to to accept the pr. So um we respond to the author, um but they do not have any further interaction with us and yeah. I agree with you. We cannot make everyone happy, but we can try our best to make most of them happen. I think and yeah.

F

Let me try my best to to review prs from the community and yeah.

A

The idea is not to accept everything. The idea is to provide feedback to everything so once someone creates a pr and tries to contribute to to feel welcome and to contribute again, so we can have that constant cycle of contributors coming back because, let's be honest, contributing to open source projects with code can be super scary uh by few reasons, and, for example, is you're not always sure your code is good enough and it will be liked into your project.

A

So we have to be super, has a welcoming and and get folks into into the speed with the project.

A

So, if you like, I can I can try to get a uh from other projects like how their process looks like and what are the responsibilities of the maintainers in terms of pr reviewing, and we can formalize that, if you, if you want, I can, I can do that.

A

I've seen projects well defined pr review process uh and what to look for how to behave. What, in what time frame to behave and stuff like this? If you think that to be careful for the maintenance, I can do that.

A

But I'm not sure if we need that formalization of the process. That's that's the thing, but if it's useful for for many of us.

F

To be honest, I I'm in the maintainer of other of another project and we actually do not have any um pr uh um process for from a community, and we just see that vrs we can discuss, apply and do some review on our free time and yeah yeah. But if someone has some formal process for uh prs and yeah, I'm okay, that they can share something with us.

A

Okay, and do we do we want to have uh like a and then every week, an hour to discuss the prs and to follow up on everything or because don't get me wrong, I'm wearing my okay. It's on this side. My vmware shirt, but still this process to be encapsulated into vmware is not the best practice, so I think everyone can benefit from having like a common meeting with all the maintainers and to do a pr review for everything open or at least like the hot topics. Right now,.

D

A

F

D

I have one just comments right. I think many times review we review the ti is time consuming, I think so, certainly yeah. I I think async. Maybe is the best way to uh to to review this pr, but if so for summer, with the pr, if it leads the maintainer to discuss or even need involves, also to discuss, I think for those kind of pr, maybe uh to help maintain, describes together usually efficiently, but for most of others. I think the async uh maybe is better.

A

Okay, then, how you see the async.

D

Of the pr yeah leave the comments on the pr and uh discussing that prx, and also everyone can view that right, it is visible.

B

This should be like this. I mean the comments should be in the pr so like to the pr requester. There should be comments, and if and if we do it like, like you you're doing it now, every week um we should have after a few weeks, we should not have come prs without comments or feedback to the to the person that does provide a pr or did provide a pr.

A

So um can we sort it like, you know the commands we have. I think.

A

Okay, there's one without the command, but okay.

A

Okay, shall we give it a try like a week or two try to communicate more into the deaf channel about prs.

A

I can I can take some role in this one, but uh it's part of the maintainers tasks to do this triaging and and to watch the list and and to command, and can we accept it? Like is a general rule, new pr, no commands, I'm gonna comment or I'm gonna, I'm gonna attack someone who understands that area of the code that can comment or something like that.

A

So we can drive that topic to to be more welcoming to to the folks and they can see that actually, someone is looking at their stuff and their effort and their time spent is not like wasted and yeah, because if we, if we take two weeks to write a command, uh it's highly possible that the creator of the pr won't come back and fix whatever we asked them to fix yeah. So if we do that review in one two days, they'll be super pumped and they'll come back and fix whatever we wanna.

A

But if that takes like two weeks, yeah screw that I won't fix that. Obviously nobody gives a I'm sorry, my language, about what I've done so.

A

C

A

That's my take and I've seen that in other projects as well. So that's how people work yeah if you're not like, if you're not getting respond shortly, the chances of coming back is super super low. So, let's, let's agree on because you are all assigned to the maintainers on every new pr created, so everyone is getting notification in their email.

A

So please take a look and if you can respond and spend some some time of your day, I would suggest everyone if everyone can add, like a 30 minutes of your day, into your calendar, to do uh pr triaging on your own time, like prior lunch or something that you help a lot.

A

I I have that in my in my calendar uh every day I have 30 minutes to go through all the pr's into the community and the website repo and some other repos, and to check if I have to comment something and not to rely on mail notifications and that by by the way, it's a good practice uh for for every requiring job anyway. Right so, and that's I've seen that in some of the governance docs that maintainers are by accepting the maintainer role, they should do that. It's part of their role, that's it so.

A

The the governance suggests that you have that spare time every day, three minutes not more, maybe 15, um to review prs and to make sure that we drive all the topics.

A

What do you think.

A

I don't want to force you to do stuff, I'm just suggesting good practices that I've seen in other projects, and I apply in my my work.

A

B

Yeah, so I mean I, I definitely agree we should. We should work more with pr, because this is, uh you know how we get people uh contributing to harbor and how this project can continue to grow.

A

Okay, I'll take a note on my side that I'm going to take a look at improvements in the governance, so people, if they get lost uh they can they can refer back to to the governance and uh and and get ideas what to do.

A

Okay, we have five more minutes so.

B

Let's get with the other stuff with him um yeah, so the other point I would like to it's: it's related with two, maybe two prs as well: um it's the the topic of the cen hub, so the cncf has to send up a subscription, and I saw that xenapp is already used in in harbor or was used in the past, and my question is: should we use it or prefer it in comparison to to project so because, right now there is the project.

B

Management is completely not part of the cncf, and this is not actually compliant with the cncf rules that it's not part of the cncf. So it should be somehow transparently part of the cncft project stuff and the options that we have here is the projects within github, which is a possible solution, or we can continue using zenhub, which was also used. I think in the past, and the question is what happened with the zenhub? Why is not used anymore and why there are projects that are not used anymore?

B

So how do we track the progress of the project or how is it tracked currently in the cncf scope and not within vmware.

A

I can answer. I can partially answer this, one, not sure about the zen hub, but um I think roger right now works on how the okay, how to put this one vmware, took the role of project management, not to be like a vmware specific. There are some specifics for vmware, of course, but in general roger is the pm assigned from from vmware to be a pm to the project and representing the community and the cncf uh interests.

A

So uh I think that question is: why isn't hub? uh No, why not uh anymore, I think roger, can answer that in details, uh but I think we have on the projects and the roadmap. I think they're on the on the projects. Isn't it or I'm wrong.

A

So yeah yeah yeah.

D

Yeah, it did not get it updated. I think roddy is working on this.

A

Yep, so I'm not sure, maybe I'm not answering uh video's question, but um there were some issues in the past which were more vmware rented and I think right now, roger is more concentrated on bringing the gap between the cncf interests and the project interests and the vmware so to be less vmware-ish.

A

So if you have some some questions about project management, I think he's the right uh person to answer them.

A

uh Of course, if someone from the uh vmware china team can answer them right now, but I think that's the overall status right now, um so that's that's how it should should be done.

B

Okay, so the send up is is not an option anymore and because I.

A

B

It it was there in the past. That's why.

A

I, to be honest- I I don't know but uh and why the project management decided not to go with the help anymore. I don't know uh maybe alex, can ask.

F

We use we use the zen hub to organize the ipad. I mean the the the usual impact to organize other engineering tasks. So if you click on the thing hub, you can see all the effect that associated with some anchor features.

B

Yeah, I see I see stuff in zenhub. Still, you know the old one.

B

Four years ago it was used.

A

And do you believe uh that help is better from the.

B

A

B

Projects or I I don't know, I'm just uh trying to find a way how we can organize and streamline this. So if we can live with with github's project functionality, let it be this way. I mean it's, it's a gradual progress. So if this is, if github is enough, let's stick to github and if it's not enough, we can update to zenhub, I mean xenapp looks more complicated, but I'm not sure if it's really needed. You know we should really start slow. If, if, if you want to embrace this striking.

A

B

Want me to set up a chat with roger about that will be, would be interesting, yeah to see what was his opinion about this yeah. Anyone else.

A

Interested in this topic, I can because I'm super interested uh in this topic as well.

A

And how we can improve it overall and to be less vmware dependent and more community oriented in general.

A

Okay, so it's me and the team then okay.

A

Okay, next one.

B

Oh, the last one um yeah the pull request review. I think I have a few few. I have a few pull requests that I think we could approve quickly because it's like typos and bug fixed and not tight box buckets mostly typos, like in swagger and a few things that we should take a look at, and I would like to I will post it in the slack channel so that everyone can take a look at those uh because it would take too long um and then that's that's it from from my science.

B

My points there is like deprecation policy for harbor discussions. Next.

A

Yeah, that's that's something that uh it's circulating for a while um and just to yeah. Let's do it async because we're on the hour actually one minute after the hour um that thing came out from um an informed decision to uh announce the uh notary, deprecation and came out. We don't have an actual working policy, duplication policy and process for deprecating features.

A

So if you, if everyone can take a look at this one and give ideas um what we can do, there will be great because uh we need we need a policy. For example in like two releases: we're gonna do this and we have to announce this and that- and it's not only procedural but also technical process, how we can.

A

Deprecate something whatever it is documentation to functionality.

A

So please take a look. It's a request for everyone. um I'm gonna paste it in the chat. If I can find the chat.

A

All right so with that, uh thank you, everyone it was.

A

uh It was very informative meeting today. I hope uh we have some takeaways and hey with him nice shirt. What are you gonna? Thank you. Well, just uh yeah, that's a good thing just to share we. We we give it away around 100 average t-shirts or 150. I I have to check the numbers, but they were like one huge box and we give away around 50 utility keys and around 300 stickers.

A

So I think we did a good job with the swag and people were like super happy with this rack uh and one last thing: I'm sorry.

A

Cfp closes in two days. Anyone uh planning I know vadim, is planning to submit something. Yes,.

B

I'm playing a bit something.

A

Okay, anyone else planning to submit.

F

And we have to, we have two cfps.

F

We will submit them.

B

Maybe tomorrow the maintainers track right.

F

No, no, no, we just have some ideas about harvard or rock harbor. Okay,.

A

Do you need do you need some help with this one.

F

uh We did some internal energy review with owner.

A

Can you, okay, so fast again, very great yeah? uh By the way? Can you can you add that to the status updates over here? So we can keep track on this one. We have two more days, so I'm I won't submit anything for for harbor. I have to submit other stuff so, uh but if you, if you need my help uh and my assistance with this one, I think we can do again the kiosk and this time we should not miss the chance to be on the keynote uh project update.

A

So everyone keep an eye on this mail. It's a it's an email from events at cncf and they ask the maintainers if we want to be on the project updates panel on the keynotes on the last day- and I think that was was a huge miss for us that we were not there with two five uh release, because there were other projects with super minor releases and changes compared to what we did in t5.

A

So I think I think, for the uh north american one we can be there and that will be super helpful for the project and for the.

B

Community as well, and also who is going to be in person there should we have a maintainers track in person as well. In detroit.

A

uh The maintainers track cfps are filed in a bit later: they're, not processed they're, not part of the uh like at the regular cfps.

A

uh So that's actually depends on on on the team who want who can and who wants to travel to to detroit I'll do my best to travel, so I can be there in person for many stuff, like even the the harbor kiosk again. So if you can get as much as possible, maintainers on site would be great yeah. That would be good yeah.

A

uh Do you, folks from china? uh It's a question for original question for china? Are you uh okay, with traveling for end of the year, for detroit.

F

Oh, I don't know, but maybe it depends on the government's rules. I think.

A

I'm sorry, can you repeat that.

F

I I don't know, I think it depends on the common use.

A

Right now it's more uh regional uh regulations than vmware regulations, because if you're right now, I can yeah I'll talk about vmware a bit um if you're, if you're a speaker you're allowed to travel, uh but then, if your country doesn't allow or the receiving country doesn't allow you to get in that's another round of problems, but I'm just asking in general. Are you willing to travel and are you okay to to dealing with travels to us at that point?.

A

But okay, uh maybe it's too early for the maintainers to discuss that. uh So please, if you, if you, can uh drop the topics for uh for your talks, so we can keep an eye on this one and we know how how that will proceed.

A

Okay, thank you. Thank you. Everyone I just.

C

Everyone nice meeting today.

A

And see you in two weeks, bye, bye, bye, everyone.