2024-06-27 Meeting notes
- Feanil Patel
- Kyle McCormick
Date
Jun 27, 2024
Participants
@Feanil Patel
Previous TODOs
Discussion topics
Item | Presenter | Notes |
---|
Item | Presenter | Notes |
---|---|---|
edx-platform Maintenance Models | @Feanil Patel |
|
Swapping Maintenance Meetings | @Feanil Patel |
|
|
|
|
Teams in Github as a communication form |
|
|
Next 2 Meetings | @Feanil Patel |
|
frontend-app-learning Hand-off | @Feanil Patel |
|
frontend-app-ecommerce maintenance? | @Adolfo Brandes |
|
Repo ownership clarification | @Robert Raposa |
|
Adding Catalogs for things that don’t have owners |
| From @Chintan Joshi I am creating catalog files for such repos which have no interest and no info etc I've finished PR for priority 1 on going work for priority 2 and should be done by Monday... and priority 3-4 I'll pickup next week |
Action items@
Decisions
- All teams that are maintainers should be public.
Recording and Transcripts
Recording: https://drive.google.com/file/d/1UB9vjSTQbDv2F8-9ky7SXNgLdVov_Pbu/view?usp=sharing
edx-platform Maintenance Sub-Group (2024-06-27 09:03 GMT-4) - Transcript
Attendees
Adolfo Brandes, Awais Qureshi, Chintan Joshi, Feanil Patel, Feanil Patel's Presentation, Jeremy Ristau, Robert Raposa, Tim Krones
Transcript
This editable transcript was computer generated and might contain errors. People can also change the text after it was created.
Feanil Patel: All right, everybody Welcome to the edx platform maintenance meeting. yeah, pal won't be joining us because he is flying over the Atlantic and apparently doesn't have internet while on an airplane over the ocean.
Feanil Patel: But we'll miss him. And hopefully he got some sleep on that airplane. in the meantime, I think there's two big things to discuss and then anything else people might have but Does anybody have anything else to add to the list before I get going?
Robert Raposa: I don't know where we want it this meaning or whatever. This is just a minor request of getting a link to the ownership spreadsheet that you use. Either in the wiki wikidak or as its own link on the slack Channel. is it but
Feanil Patel: Yeah, do you want to just go ahead and here's the link? Why don't you go ahead and add it to the slack Channel?
Robert Raposa: That would be great. And if there's other But there are other ducks that point to that that would be a better.
Feanil Patel: Yeah.
Robert Raposa: Place it?
Feanil Patel: No, I think it's everything.
Robert Raposa: wanted to
Feanil Patel: Yeah, it's the thing that everything is pointing to you. And for…
Robert Raposa: I've got it.
Feanil Patel: what it's worth that itself is simply a snapshot of the catalog info data and also sort of working data for who I need to talk to to make sure that things need to get updated so It's a working document more than a artifact that you should say as the economical like truth of anything.
Robert Raposa: Yeah.
Robert Raposa: yeah, I was just curious from the cookie cutter perspective because I don't think there's a catalog by all and so this is
Feanil Patel: Yeah, yeah, if there's no catalog file, yeah, then there's no owner in here. There might be a potential owner, but I don't think anybody's listed at the moment. Yeah.
Robert Raposa: yeah.
Feanil Patel: And I'd like to start I think for moving forward. Any places where we're we own that but we don't want to we should follow the maintenance. Project process so with the front end app authoring. I think we've talked about that. I'm gonna talk about that a little bit. Later, but for edx platform real quick. I did finally start that document for sort of potential ways. We could start splitting the sort of the high surface area of work that is maintaining at X platform.
Feanil Patel: There may be other models that we should add to that and then probably discuss in a couple of weeks. What our options are and what actions we might want to take sort of towards those options?
Robert Raposa: Do you plan on and you want to present a high level of what's in the document now or?
Feanil Patel: Sure, the document is pretty simple, which it's got a template and then it's got a potential model. So I assume that people who want other potential models in here would add them. With whatever pros and cause they can think of but then probably we'll have a review cycle where everybody goes in and be like, the problem with this is blah blah blah or this is great because blah blah and we'll keep updating these lists. and then come to hopefully a decision at the end for what that model should be on top of having a repo maintainer
Robert Raposa: You got it. So this is about getting more groups,…
Feanil Patel: is that it's
Robert Raposa: but not necessarily about. What their responsibilities are and…
Feanil Patel: So yeah,…
Robert Raposa: what or is that in here?
Feanil Patel: I mean this is how we were Distributing responsibilities you're thinking what are the defined responsibilities which I think are kind of an extension of the existing maintenance responsibilities for a repo just slightly further down the file tree. Right. So making sure maintenance upgrades happen in that area.
Robert Raposa: Okay.
Feanil Patel: If I'm like, hey, we are doing the Django upgrade. Please review your own edx platform area for any warnings that need to be fixed before we can ship Django five to that is now your responsibility to get over the line. And as the maintenance we want to be helpful for that is a list of what warnings exist across the entire system. Let's see if we can do some work to produce that list for people.
Feanil Patel: But we may not have capacity for that in which case be like go look at the warnings in your space and go look at the change log.
Feanil Patel: Who owns the determination writing of issues related to specific owned code or that costs around? I think anything where it crosses boundaries like that comes back up to the maintenance of the platform to be it crosses these boundaries these two are the owners. Let's all three of us get into a room and figure out who's doing what and who has capacity and so resolution of that stuff. I think sort of bubbles up to the repo maintainer. In the same way as things that cross repo boundaries might bubble up to the maintenance working group.
00:05:00
Feanil Patel: That's my thought. I could have everything bubble up to the maintenance working group which mechanically is different but in terms of reality is the same people deciding the same stuff.
Robert Raposa: And then for a kicking off. Who are the owners? are we thinking it's go back to Jeremy's spreadsheet of here's all the old two you teams. which ones did we actually want to keep it?
Feanil Patel: is that
Robert Raposa: I think it's not gonna be what it was but just hey,…
Feanil Patel: Yeah.
Robert Raposa: here who was owning things do they think they're still owning it too. They want to still own it. Do they not want to
Feanil Patel: I would love to answer that question. Once we've decided what model of owning we're doing because the answer might be different based on how we split the platform. So
Robert Raposa: Got it. And what are you calling a model? I don't understand if the model because you said the responsibilities are just a smaller scope of Whatever the maintenance is. So is that…
Feanil Patel: right
Robert Raposa: what the model is which sounds simple and reasonable or is the model just here's how we're going to implement knowing…
Feanil Patel:
Robert Raposa: who owns?
Feanil Patel: here's how we're going to implement. How we split up the platform? Is what I was thinking of is the models but maybe there are other ways of viewing this which is like we have this entire platform one way to maintain it is to split it up. So that many people have subsets of ownership and there are many ways of even doing that another way to split it is to just say the platform is actually three things CMS and everything else and we just need three maintenance that are sort of sub maintenance of the main one. And their responsible for whatever.
Feanil Patel: Another way to split it might be like all of the Django apps need individual owners and then everything else is one owner is yet another there are many ways to cut this thing and…
Robert Raposa: Okay.
Feanil Patel: depending on how we cut it. We may want to choose who there's no matter what we do. I think we're gonna have more people responsible for code. And so the answers to questions, how do we decide who those people are and how to distribute responsibility to them or stay the same? Whether or not we'd like so I think that's like stuff we can defer until we know what model we're doing to have so that we can for more precise outcome
Robert Raposa: This yep. So your last answer was very clear on helping me understand what you actually mean by a model and what the different models are and…
Feanil Patel: Yeah.
Robert Raposa: I'm wondering even if this is an ADR with rejected Alternatives here are some other models that were discussed and that we don't plan on doing or something of that nature.
Feanil Patel: It's so the last conversation we had about this was we want to start writing down all the possible models because I don't think we've decided on one yet. So I think Rather than going with rejected Alternatives just put them in here and…
Robert Raposa: Yep.
Feanil Patel: then we will like that is later.
Robert Raposa: Yeah, that sounds fine. I didn't know if you were making a proposal as well. and not that that would be decision made…
Feanil Patel: No, no.
Robert Raposa: but even starting ADR I meant but if it's just listing all the potential models now, I understand you've got one example here.
Feanil Patel: Yeah. Yeah.
Robert Raposa: There's other examples we need to get in here and now I'm clear on.
Feanil Patel: if I put this one here because I think this one probably makes the most sense to me or after five minutes of thinking but splitting it up as LMS and CMS and having a high level team that owns each of them. Maybe is a better model because then there's less finagling to do in that we'd sort of distribute the managing underneath of that to somebody potentially.
Robert Raposa: And it could be a hybrid that starts with that and then layers on specific apps that make sense,…
Feanil Patel: Yeah.
Robert Raposa: .
Feanil Patel: Right, but I want to like a part of the discussion we will have once we have all the models as this one means that there are seven layers of hierarchy. Is that valuable to us? This one means that it's extremely flat. But now we need to be communicating to 300 different Django app owners. Is that valuable to us? those are the kinds of questions that will come up once we have some other models that we can compare and contrast against
00:10:00
Robert Raposa: but that's
Feanil Patel: And so yeah, I invite you to add more models, but also maybe change that top paragraph Robert because I think Maybe that is not enough information and maybe you can add more information based on your understanding.
Robert Raposa: I mean, I will admit I was looking at titles and not reading closely. So the top paragraph might have answered all my questions. I don't actually know yet.
Feanil Patel: Okay, I'm gonna add you guys as editors but by default the sheet is comment,…
Robert Raposa: Okay.
Feanil Patel: but not at it. I'm happy to give anybody at it. because it's gonna get posted on the public Wiki. I didn't want to just give blanket edit access
Robert Raposa: should we just add just quickly though your brainstorm even if it's as bullet points post that or titles of here's other.
Feanil Patel: Yeah.
Robert Raposa: that's CMS everything else
Jeremy Ristau: can I just ask why we need to come up with the new way because we're already doing this, there's already an edx platform code owners file and it already has owners in it for a different directories inside of edx platform.
Feanil Patel: I think we've tried that a couple of different times and nobody has ever enforced it or made sure it's accurate and staying up to date. So one of the options could be let's double down on the thing we have but actually us as the repo maintenance will own keeping it up to date and making sure everything has owners that we want to distribute ownership of
Feanil Patel: I don't think code owners as a tool is a bad idea, but I don't think how are we thinking about the strategy for managing that it's platform maintenance?
Jeremy Ristau: yeah, absolutely, but in terms of Divvying up ownership with labels like code AS doing that for Miles. Yeah.
Feanil Patel: Yeah, yeah. Absolutely. I think that no matter what we do that is probably a mechanism. We will want to use because one of the other nice things is you can go to a file and…
Jeremy Ristau: right
Feanil Patel: GitHub will tell you who the code owner of that file is also so there's some really nice to have with
Jeremy Ristau: Right. Yeah.
Jeremy Ristau: so to me, this is less I think what you're pitching is the tactical implementation But it isn't really a model to me like the strategy is the model. Are we going to go Seven layers deep is it yes subset of people.
Feanil Patel: right
Jeremy Ristau: Is it going to be open to anyone because right now apertures in there? But they're not in edx platform maintainer.
Feanil Patel: right
Jeremy Ristau: So those are the real questions that I have about maintenance as opposed to are we going to use code owners?
Feanil Patel: Yeah.
Jeremy Ristau: We're gonna use catalyt like that feels like way local and…
Feanil Patel: Right. Yeah. Yeah.
Jeremy Ristau: I'm wondering if starting at this level is getting lost in the trees.
Robert Raposa: And it may just be that your first model. Needs to be just Rewritten a little bit to match. it's more for any path or is it maintenance per Django app or what? level are we going to be cutting stuff down in terms of the signing maintenance, but not
Jeremy Ristau: I mean As an example in codonors we go down to a specific file for tnl…
Robert Raposa: And we're impoters.
Feanil Patel: Yeah.
Jeremy Ristau: because there is a specific file that needs to be socks compliant reviewed by someone inside of T&L for any changes.
Feanil Patel: right
Jeremy Ristau: So I don't think tnl should own the top level directory that that file is in but the only that specific file is very very important to you.
Feanil Patel: right
Jeremy Ristau: So I wouldn't want to abandon the ability to assign direct ownership of specific locations inside of edx platform whether we go with code owners or…
Feanil Patel: Yeah.
Jeremy Ristau: not. So I hope we can divorce the maintenance from the assignment of ownership or something like the maintenance strategy from assignment of ownership. Who's involved in maintenance is a question. I think is it everyone who owns something in the monolith? Or is it just people who are in the maintenance group? That a big question…
Feanil Patel: right
Jeremy Ristau: because I don't know if I need to go talk to Kelly and say Kelly sorry, you're not gonna own credentials anymore because it's an edx platform and only people who are maintenance of edx platform can own that so I think those are some open questions a little bigger than I have.
00:15:00
Feanil Patel: Yeah. yeah, I mean I think along those lines I
Feanil Patel: one of the things when you said that sort of concerns me is people who own things who are not. either helping or have some way of ensuring that maintenance of that's content is being done pivotal level that we want and at the speed and SLA that we want. So it may be that Kelly owns it but you is supplying sufficient maintenance maintenance out like maintenance people to edx platform that they don't worry for all of the maintenance across all of the two you related things these seven people are going to handle all of it. So you can communicate with them for maintenance requests. As opposed to all seven teams individually and that's a process question that should go in here. the potential model is there are I think to me maintenance separate from ownership is a model that should go in here with a description of how that would work because that's one of the ways we could do it.
Jeremy Ristau: Do you think there could be a section added to this that's like problems trying to be solved?
Feanil Patel: yeah, I mean, I think that this top here should have I think this is
Jeremy Ristau: Or use cases to consider or that kind of like the running list of things the constraints I guess is maybe the right word that the models need to fulfill.
Feanil Patel: yeah, please add things here. Yeah. so
Jeremy Ristau: Yeah.
Feanil Patel: Yeah, I think that's a super useful thing. I figured this was going to be one of these things where we're gonna iterate on it a bunch because we have nothing at the moment that we're happy with.
Feanil Patel: And it would be good and to understand what our concerns shared constraints are.
Jeremy Ristau: Yeah. the socks compliance thing, you…
Feanil Patel: right right
Jeremy Ristau: that's a very specific both in the reason it exists and in the place that we need to Target for it,…
Feanil Patel: Yeah.
Jeremy Ristau: that's a random one, but super important one group of people.
Robert Raposa: right, but that one also, I mean, it's great because it's a real constraint and…
Jeremy Ristau: right
Feanil Patel: that
Robert Raposa: there might be many ways to solve that It into its own repo.
Jeremy Ristau: right
Feanil Patel: yeah.
Robert Raposa: Or whatever, there's so many other ways.
Jeremy Ristau: Yeah, right exactly. Yeah.
Feanil Patel: We could be adding this thing needs to shift to a plugin is a priority that we have as maintenance now because that simplifies our work.
Robert Raposa: right exactly
Jeremy Ristau: Yeah.
Jeremy Ristau: We'll turn the student model into a plugin. That'll be fine.
Feanil Patel: I mean
Feanil Patel: I think you could do it I believe in you.
Jeremy Ristau: Thanks. Sounds like roadmap 3:30. I'm there.
Feanil Patel: Yeah, how that helps to use bottom line on clear. Okay. Yeah,…
Jeremy Ristau: Yeah, right exactly. Yeah.
Feanil Patel: so please add more things to it. I think we'll just revisit it. every maintenance meeting a little bit so that we can keep iterating on it, but I would love for people to spend time in between but
Feanil Patel: and then the only other thing I had today was
Feanil Patel: Some child art things enjoyed and…
Robert Raposa: On the other thing again. I get
Feanil Patel: yeah, that's Yeah.
Jeremy Ristau: good.
Jeremy Ristau: Shades of blue that's very impressive.
Feanil Patel: Yeah. It's very wet.
Jeremy Ristau: the ocean
Feanil Patel: Translation. Yeah, it's on brand. Yeah, the only thing was I think I was thinking about swapping the nine o'clock and 9:30 meeting so that the full working group meeting is at nine. Does that conflict with things for you guys is like I'm trying to remember why we had it this way other than.
Robert Raposa: I mean, you're talking about swapping meetings that were in both of so,…
Feanil Patel: Yeah.
Robert Raposa: I think we just had it this way because the other one came first and…
Feanil Patel: Okay. Yeah, okay.
Robert Raposa: it was on the calendar first and then we added one spirit. Swift swapping sounds fun.
Jeremy Ristau: Mm-hmm
Feanil Patel: I kept noticing that we keep talking about General stuff and then I either have to rehash it or we have to hold on it and then do it later and I think
00:20:00
Jeremy Ristau: I think the reason that we have it this way is we originally had edx platform after but then the time change happened and it went from 8:30 to 9:30.
Feanil Patel: rare Right, right.
Jeremy Ristau: And so if nine it will be at 8:00 am just the note.
Feanil Patel: Okay, cool.
Feanil Patel: As long as the one fall I think the edx platform following the other one will make more sense just to my brain,…
Jeremy Ristau: Yeah, yeah, that's
Feanil Patel: so Yeah,…
Jeremy Ristau: I think we had that original and then the time. Yeah.
Feanil Patel: and then with the time change here to set yeah.
Robert Raposa: So maybe just knowing that people that during the time change both are going to move.
Feanil Patel: Both are gonna move earlier, which will hopefully continue to stay away from your other meetings because I think that's what happened. Was that that 10 o'clock to 10:30. You have a conflict.
Feanil Patel: All right, we have five minutes. so we can kind of iterate on this for a little while since we've got five minutes before the other meeting and then just hop into that meeting as people arrive.
Robert Raposa: please
Robert Raposa: So Jeremy you are trying to make a point about. ownership versus maintenance ship and…
Jeremy Ristau: it doesn't sound.
Robert Raposa: I don't know what
Robert Raposa: yeah, what exactly you wanted day and…
Feanil Patel: Yeah.
Robert Raposa: how to make that a constraint or how to add to maybe the first bullet point to
Jeremy Ristau: I added a second bullet point which is essentially that it's more the problem.
Feanil Patel: Yeah.
Jeremy Ristau: Or maybe it isn't a problem.
Feanil Patel: I guess yeah,…
Jeremy Ristau: It's the current.
Feanil Patel: I was gonna say Yeah. the current state of the problem is that a bunch of people can make changes to edx platform and let's say very few of them are helping maintain edx platform.
Feanil Patel: I don't know if the owner might be the wrong word because honestly a bunch of CC's can make changes to parts of edx platform and are also not currently owning things. So it may just be that we currently have maintenance and people who can change code and they should all maintain a
Jeremy Ristau: no, And we also have people in the code owners file that teams are marked as owners of locations inside of edx platform that are not in the maintenance working group or…
Feanil Patel: Okay, right, right. Got it. right
Jeremy Ristau: in the list of people who would maintain at X platforms.
Feanil Patel: Yeah.
Jeremy Ristau: apertures one tnl's another they're not named teams that maintain the code.
Robert Raposa: And I guess yeah part.
Jeremy Ristau: and is that to you owning portions of edx platform and dealing with it underneath and then somehow we have a 2u team that we slap on directory there's Logistics that come from that problem depending on…
Feanil Patel: right
Jeremy Ristau: what the solution is, but it …
Feanil Patel: right Yeah, yeah.
Jeremy Ristau: it's a constraint that we need to
Robert Raposa: Yeah, and I think there's two sides to this one is we've got teams in the code on our file and we don't have a very good definition of what it means when you are in the file, …
Jeremy Ristau: yeah.
Robert Raposa: what your personalities are and on the flip side we have teams that are in that file probably for Are those reasons listed here is constraints for. some of Kelly's team She did it because she wanted to get notifications on all changes in a particular area. And is that Kim?
Feanil Patel: right Is that okay? Yeah.
Robert Raposa: Will and anyone is solve that problem in that way if they want to can only certain people solve it…
Feanil Patel: Right, right.
Robert Raposa: because they are now and…
Jeremy Ristau: Yeah.
Robert Raposa: owner and what is an owner mean then or…
Feanil Patel: If it becomes it,…
Robert Raposa: Tanner?
Feanil Patel: right if we use it as our canonical way of tracking ownership. We lose the ability for people to watch parts of the platform they care about Without being responsible for them.
Robert Raposa: and unless we are able to comment the code on her file and…
Feanil Patel: Yeah, yeah, right.
Robert Raposa: make sure they're like, okay, this is purely for notification purposes you
Jeremy Ristau: Yeah, although I believe Kelly might have wanted it for more aggressive reasons than notification and…
Feanil Patel: right
Robert Raposa: correct. right
00:25:00
Jeremy Ristau: possibly more like gating but yeah, yeah, I think there are levels of that.
Feanil Patel: Yeah, yeah. right
Jeremy Ristau: Right and we're gonna need to figure out how to support things like engage stakeholders or internal owners inside of one particular Community member or…
Feanil Patel: Yeah.
Jeremy Ristau: And maybe the answer is something like will they become core contributors that ex platform because they're not maintenance or I don't really know…
Feanil Patel: right
Jeremy Ristau: what the answer is but it's a case that we're gonna need to solve for sure.
Feanil Patel: We're definitely gonna have people who can make changes of the code who are not responsible for maintaining it. I think the thing that we want to figure out is…
Jeremy Ristau: right
Feanil Patel: it may be helpful just to think about that as a separate issue and just say we always have people who can make changes to the code who are not maintenance of the code, but If you want to be gating a piece of code, you need to either be the maintainer or Supply sufficient maintenance like you need to complete the maintenance SLA whether you do it yourself or have somebody else do it like…
Jeremy Ristau: Yep, and…
Feanil Patel: if you like.
Jeremy Ristau: then that's really a question of are we talking about to you? For this thing and…
Feanil Patel: right
Jeremy Ristau: we have to figure it all out internally and we just slap to you on a bunch of directories or is it team by team? We make that clear to the community and then it's more decentralized and yeah
Feanil Patel: I think it's a question that I don't know if it goes in here, or is the question that you need to ask sort of itself emotionally is what level of Team level engagement with the open edx community. you wants to be doing right and this is also related to the core contributor stuff that we've been talking about separately. if you work for contributors, you are part of the community and therefore a certain level of responding to questions on tickets and forums is expected not a lot. I still just finished reviewing a PR that I was tagged on two months ago. So it's not just a two But certainly I take responsibility for the fact that I took so long to do it and able to be reached.
Feanil Patel: and so I think there's a question of
Feanil Patel: The 2u teams internally are not engaging as much with the community and they're sort of like an interfacing team. Then maybe that team is the quote unquote owner until you figures out internally what that means for you guys, but then that team can be sort of communicated with and can respond to things and that's one model that you guys could have internally.
Feanil Patel: All right.
Jeremy Ristau: I hear that a lot and I wonder do people ever so We have the same problem right teams internally. Need to work with other people internally or they need to get something from the community. When we need something,…
Feanil Patel: Yeah.
Jeremy Ristau: we generally comment in a PR and go to Slack. Does that happen?
Feanil Patel: right
Jeremy Ristau: When people are trying to get in touch with people to you, I don't know but people reach out to me, but that's a different. thing
Feanil Patel: I think part of the problem historically has been there's no obvious mapping of GitHub to slack and not everybody is on the open edx Slack and they can't reach to you to communicate with everybody. So there's some of those communication. gaps exist in terms of I think the question you're asking was like if there's an external person that needs to reach it to you person. How do they do that today? Do they go on the pr and then go in Slack?
Jeremy Ristau: Today, I think they just like what I usually see is in the pr. team And then a week later team and…
Feanil Patel: right
Jeremy Ristau: then a week later team and the same. Action over and over if it doesn't warrant. if it's not getting the result you want you would think you'd take a different path after a while and are there other paths that open edx computers can take I don't know.
Feanil Patel: I think that's a problem is that there aren't clearly defined other paths that they can take There's no obvious way to reach. those teams in slack because I'm sure that for each of those teams. There's a team slack channel to use slack and it would be really easy to find that and go be like, hey, please help me but that doesn't exist on the external side. So it is harder.
Jeremy Ristau: that I mean does it exist for open craft or raccoon gang or edley? Or how do other companies interact with each other?
Feanil Patel: For the other companies for the most part they have named human maintenance for repos, and those people are on slack and can be reached directly. and they are on the discussion forums and can be reached there and they're yeah.
00:30:00
Jeremy Ristau: Okay, so they just identify a human being instead of a team.
Feanil Patel: Yeah.
Feanil Patel: Which I think becomes a lot easier at that point, but I think I don't know if actually and that may be a thing we can do on the GitHub side. Which is a we could.
Feanil Patel: potentially make teams public
Jeremy Ristau: is that true? I mean committers front end. cookie policy Banner maintenance domain translations maximum engineering backstage containers
Feanil Patel: Yeah. Yeah,…
Jeremy Ristau: those are
Feanil Patel: I mean there are definitely teams for certain things. But I think most of the other repos 80% of the things you just named are people that accents. because yeah,…
Jeremy Ristau: right, but who would know that?
Feanil Patel: yeah, so there may be a question of can we
Feanil Patel: I don't know. So there's two parts of this My one sassy answer is yes, but if you actually just pay attention to your GitHub notifications, you can just respond to them without them having to seek seven different places to communicate with you, which I think is a useful Baseline expectation.
Jeremy Ristau: Mm-hmm
Feanil Patel: the sort of further answers it would be useful to be able to see The people on those teams, so I wonder if we need to make all of the teams that are used for ownership public teams. If they are not already the link for public teams. You can see their members and then perhaps you could do this further. following but I don't really condone a person having to chase down people to have their stuff reviewed. I feel like if you are the team that is the maintainer you should be responding to PRS on the place that you're maintainer.
Jeremy Ristau: So as an example of teaching and learning team getting ping, but the team was switched to two utml. You…
Feanil Patel: right
Jeremy Ristau: that's a tough one to communicate out to everyone because they just know the old user the old.
Feanil Patel: Right, right.
Jeremy Ristau: Still sitting around but there's not an escalation path when you keep using that handle, but it doesn't have an effect. And so I'm wondering Is there a need for some way of escalating Beyond just continuing to Ping in GitHub? I don't know. because I Won't be able to reach…
Feanil Patel: I think so. I mean Tim has been.
Jeremy Ristau: who they need to reach, …
Feanil Patel: Yeah.
Jeremy Ristau: ideally. Yeah, we would all just live and GitHub and we'd read all 6,000 GitHub emails. We get a day and whatnot. But sometimes you need something done and you want to be able to reach somebody.
Jeremy Ristau: And I don't know how other community members do it.
Feanil Patel: Okay.
Jeremy Ristau: So I'm just wondering if there's already a precedent or something.
Feanil Patel: Yeah, Tim, I'd be curious for your input on this which is sort of When teams aren't responding on PR's do you feel like you have an escalation path today? I know you've started using slack more to sort of start communicating when things are trailing, but I wonder I mean one answer to this might be just raising it with this group feels like the right place to help escalate those things because that will also tell us which maintainers are not being good maintenance which tells us if we need to find different maintenance.
Jeremy Ristau: Yep.
Tim Krones: Yeah, I think It's a bit of everything. It used to be that so when we initially introduced the ospr management role, we had an agreement sort of with Ned and Kelly to you that we could come to them. If we didn't get a response from code owners or maintenance after I believe weeks it was and so, Michelle and I would just kind of collect PRS that get stuck and then every once in a while we would pin these two and just kind of give them a list of PRS that were stuck and that worked reasonably. I would say as an escalation path, but I mean, that's not
Tim Krones: As far as I understand not really in place anymore. So like you said I started pinging on slack a bit more in the core contributors a channel. I think we've also used the maintenance Channel a little bit. And yeah, I think that's what we have been doing. So I've tried to sometimes look up like you said the people that are part of specific teams.
00:35:00
Tim Krones: But yeah in general I think my is to first try and ping the team and then if I don't get an answer there. Try to Ping individual people just on Slack.
Feanil Patel: yeah, but it sounds like
Jeremy Ristau: it sounds like that model that you were talking about before just like a point person and they would be like the Brother teams. Yeah.
Feanil Patel: At the very least I think. all teams that are maintenance
Feanil Patel: should be public. So you can see who the people are on a maintenance team that feels like I don't know if it's actually true, but I can write some code to make sure it's always true. So, I think that's a reasonable decision moving forward. Does anybody see anything wrong with that? This was just allow people to be able to see who the members are of a team. If it is a named maintainer of any Repository.
Jeremy Ristau: You can't do that today.
Feanil Patel: You can but teams can be public or private. So maybe that when a team it was created private and then you can't see the people who are on it.
Jeremy Ristau: Gotcha.
Feanil Patel: GitHub has the ability to hide this information. So we may have accidentally done it.
Jeremy Ristau: and catch
Feanil Patel: Go for it Tim.
Tim Krones: Yeah, one more point of confusion that I've had that I wanted to mention is that it seems that the information about teams is not properly synchronized between GitHub and backstage. So for example to you Aurora, the members that are listed in backstage and the one that I find when I go to GitHub. There's an overlap but there are different today. So I think Yeah, if we can make sure Any resources listing people have the same info? That would be great.
Feanil Patel: All Yeah, will you file that as an accent request? Forming Tim just yeah,…
Tim Krones: Sure, I can do.
Feanil Patel: that'd be awesome. That way I can follow up with it later. But yeah. I think we definitely need to do some more care and feeding of Backstage so that it can continue to be useful to people. Especially as we push more things into catalog files.
Jeremy Ristau: This is the whole downside to getting to the person level right? There's lot more to maintain that a lot more change. Yeah.
Feanil Patel: Right, right.
Feanil Patel: Yeah. Yeah more changes have to be in more places and correct more often for people to trust the system.
Jeremy Ristau: And I guess as I'm thinking about it at this very moment. Is there a reason to have people in backstage? could the two you Aurora team just have a link that points to the GitHub to you over our team. And then you don't have to manage two lists of humans.
Feanil Patel: So the backstage team should theoretically be updating from the GitHub team. It's not different. This is a book There is a synchronization…
Jeremy Ristau: I see. Okay. There is a synchronization. It just might be full.
Feanil Patel: but it might not be happening correctly for some reason. So I just want to invest like this is just literally our tooling should be better than it is, right.
Jeremy Ristau: Okay. That's good.
Jeremy Ristau: Yeah, yeah. Fantastic.
Feanil Patel: All next few meetings Kyle and I are both gonna be at the conference next week and both on PTO the week after so at least for the next platform meeting probably worth canceling for the next two weeks, but also unless somebody wants to run this meeting. I think we should cancel this one as well and pick back up on. What is that July something or other?
Feanil Patel: person
Jeremy Ristau: Yeah, I think all us people will be out next week anyway, because it's July the fourth, so
Feanil Patel: Right, so skip the 11th and then pick back up on July 18th.
Feanil Patel: All right, and I'm not hearing any concerns.
Jeremy Ristau: Maybe to talk to try and put a short-term path Record for Tim and Michelle if you want to start leveraging the maintainers working group channel for escalators.
00:40:00
Feanil Patel: Yeah.
Jeremy Ristau: I'm not gonna sign Robert up, but I'll sign myself up to temporarily be someone who can route things internally and try to escalate.
Tim Krones: That would be great.
Jeremy Ristau: but it's gonna be informal and influence based not Authority based internally, but we'll do what Ned and Kelly tried to do.
Tim Krones: I think that would be very helpful. I mean, I think judging from the spreadsheet. There's still a bunch of repositories where we haven't really found maintenance for and so for those I think it will be especially helpful.
Feanil Patel: right
Jeremy Ristau: Wait for repos that don't have maintenance that's totally different than what we were talking about. We're talking about trying to get maintainers to be more engaged and…
Tim Krones: That's right.
Feanil Patel: Responsive. Yeah.
Jeremy Ristau: PR. Yeah.
Tim Krones: okay, I guess what I had had in mind was repositories that were historically maintained by to you but
Feanil Patel: Yeah, those I think. We should use the maintenance channel for both of these but you should make it clear which it is. So I think Tim for anything's were the maintenance currently listed to you maintainer. You should escalate and tag Jeremy and Robert. I'm going to sign up Robert also on those PRS and…
Jeremy Ristau: or just Yeah.
Tim Krones: All right.
Feanil Patel: for everything else. You should still post them in WG maintenance because I think maybe can we have people who can definitely do some of those reviews and also that'll give us a better idea of which place is to prioritize finding maintenance sooner.
Tim Krones: Yeah, that sounds good.
Feanil Patel:
Tim Krones: And so about cross-coasting in the core contributors channel. Should I still do that or is that something that we only go to if we really cannot find any? Yeah. people from maintaining teams to review
Feanil Patel: If you want to I think what you should do is just forward for the ones where there are no maintenance just forward that message from maintenance into core contributors for now. That work. Yeah. Okay, cool.
Tim Krones: Yep, sounds good.
Feanil Patel: Yea