Maintenance Working Group Meeting (2024-02-29 08:41 GMT-5) - Transcript

Attendees

Adolfo Brandes, Awais Qureshi, Chintan Joshi, Fateme Khodayari, Feanil Patel, Felipe Montoya, Jeremy Ristau, Kyle McCormick, Mahendra Chaudhari, Maksim Sokolskiy, Maria Grimaldi, Michelle Philbrick, Piotr Surowiec, Robert Raposa, Sarina Canelake, Spencer Tiberi, Tim Krones, Xavier Antoviaque, Yagnesh nayi

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 for those on the recording sorry, we started a little bit late the summary of everything so far is that cc's should have merge rights on repos that shouldn't be blocked by maintenance but maintenance and cc's should be talking to each other and if you are maintenance and don't know who your cc' You can look that up on the CC's wiki page, but also you can make an action request to get a detailed list, but it is the responsibility of the maintenance to be aware of all of the people who have merge rights in their space and they are not allowed to unilaterally block CC's from merging. They should be talking to each other. cool

Feanil Patel: next is just an announcement. I just want to make sure that this gets widely seen which is that the deathstach repository is going to get removed from the organization in about a week. I think we have promised it's removal about four times over the last two years, but this time for real, so Just if that is a thing that you care about and want to retain in the opened x side of the world. Please chime in on the discussion topic there. It needs a maintainer. But otherwise the plan is for it to be archived and any organizations that are using it they can Fork it for their purposes, but

Feanil Patel: Recommend is that you guys try to use tutor and try to complain and not try and do complain about any issues that you have with development and tutor so that we can make the tutor development environment better moving forward for the dev experience.

Adolfo Brandes: If it's about developing mfps, you can complain directly to me.

Feanil Patel: There you go. And Adolfo for General complaining is it issues in the tutor? Repo? where should you complain?

Adolfo Brandes: Of an issue and any of the tutor repos if where it go this is a tutor issue open an issue there if it's Twitter MFE issue open an issue elsewhere. That's ideal. But if you just want to blurb out to complaints and…

Feanil Patel: Okay.

Adolfo Brandes: hashtag tutor.

Feanil Patel: Okay complaining this. Yeah.

Adolfo Brandes: That is also perfectly fine.

Kyle McCormick: No is the Forum instead. That's where the two remain Tanners are watching for things.

Adolfo Brandes: Yeah.

Feanil Patel: All will you just drop a link to which? Place there's the tutor subsection. in the forums and if you just drop a link for that in here for

Feanil Patel: Just like put it in there Next python 312 which we need for Redwood so that we're not running a unsupported version of python for Redwood.

Feanil Patel: The awais and the team at arbisoft has created. Draft PRS across all of our python repositories that at least turn add CI for 312 and update talks for three twelve. There's a bunch of other python classifiers need to be updated. In other work might need to be done test might be failing but at the very least those PR is now exist as a starting point. So if you are the maintainer of any of those repos, please take a look at the pull request the pull requests are all linked from the issue that I've linked. in the meeting notes, if you are not a maintainer, but would like to help and would like to sort of work your way up to being a maintainer or have capacity any repositories where there's a draft PR. You can certainly chime in and say I will take a look at getting this ready to be merged.

Feanil Patel: If you're not a maintenance, you'll have to basically make your own pull request on top of that pull request or make a copy where you're making fixes forward but coordinate by communicating on those draft PRS. About what you're doing. But yeah, if you have capacity in your teams or in your organizations or individually, please start taking a look at those, please let's get those landed if you need help or guidance to do that. Feel free to reach out to me and others in the maintenance Channel. But our goal is to get as much of the Python libraries updated so that we can start working on the core services.

Feanil Patel: questions

Feanil Patel: Okay. next

Feanil Patel: we're just talking about no 20, but it turns out that edx platform a small insignificant part of our system is still running node 16, which is not great since node 16 is already out of support the core problem there seems to be that we don't actually know what the requirements are that we need to be able to run node 18. Let alone 20. So Brian Smith from accom has started investigating sort of the root cause of that I got thank you Jeremy for all the information you provided. I think that was a good starting point, but there's no clear answer there of Here the actual blockers. So I've asked Brian to sort of dig into it further and sort of identify it as a part of that. We learned fun facts, it's not even running the latest notes 16 versions and upgrading to the latest notes 16 already breaks that X platform. So there's a couple of different layers of the onion that we are slowly peeling away here.

00:05:00

Feanil Patel: But I'm hoping that we'll have a lot so good.

Jeremy Ristau: That I thought the 16 blocker was the webpack and npm. dependency

Feanil Patel: It seems like there are things that are even before that and where the fixes haven't landed. So We're trying to dig into why the webpack npm stuff breaks and it turns out there's other things that are sort of prereqs of that. That are not clear.

Jeremy Ristau: Okay. All…

Feanil Patel: Yeah. Yeah.

Jeremy Ristau: Yeah. Yeah, I think the bomb team ran into blocker when trying to upgrade webpack and…

Feanil Patel: right

Jeremy Ristau: npm and that's as far as they got. So yeah, if you're picking it up from there and running it,…

Feanil Patel: Yeah, yeah, so we're trying to figure out sort of like…

Jeremy Ristau: that's all.

Feanil Patel: what are which packages are actually causing the issues and what and…

Jeremy Ristau: Yeah. Brew cost

Feanil Patel: where but as a part of that we learned that we can't even run the latest notes 16. So there's actually four minor for patch versions of note 16 that we haven't picked up which is not exciting to me, but we hopefully now have somebody on it that'll sort of stay on it until we can get this unblocked.

Feanil Patel: So that is moving forward. and then lastly CS comment service. Thank you Max and our raccoon game team for start doing a bunch of manual testing, but it looks like the Ruby upgrade from 30 to three one is looking good. So my plan is to merge those changes to CS comment service and inform maintenance in the next couple of days at which point I think it should be able to start running on Ruby three one instead of three zero which at least pushes us past. Needing to update anything for Redwood.

Feanil Patel: I think there's still a mango upgrade that needs to happen over here. So Max, if I make a couple more changes, would it be possible to sort of get a retest of that stuff? Because I'd love to be able to upgrade mongo and see if we can go to an even newer Ruby version before we land that's

Maksim Sokolskiy: Yeah. Yeah I have the running instance was it and I can quickly upgrade it to any new changes and allocate team.

Feanil Patel: Okay, great. I might spend another sort of half hour to hour on US see if I can get it up and running with a later Ruby version so we don't have to do this dance again for a little while. But at the very least probably we can land this and I'd love to do the Mongo upgrade and see if that is an issue share me and Robert. Them. I don't think I have to do anything that we didn't have to make any changes that are breaking so should still work with the current versions as of right now.

Feanil Patel: but I don't know sort of like what your timeline is for being able to update because in review, we currently have two separate gen files and two separate gem lock files for the two different versions of Ruby because there are sort of requirements incompatibility is between the two versions. so if we move forward at some point, I'd like to get rid of the old files, but I don't want to do that until you guys have upgraded your System that's running off of Master. So we'll probably want to coordinate that. at some point

Feanil Patel: and go for it Robert.

Robert Raposa: Yeah, yeah that was related to my question, which are there. Actual manual upgrade steps that are required and the documented How do we get?

Feanil Patel: No, no. So for the good news is the major changes. I think the only change you would have to make to your deployments.

Robert Raposa: All right.

Feanil Patel: Robert is upgrade the Ruby version and change which gem lock file you bundle install from.

00:10:00

Feanil Patel: So there's a second one and hilariously there have been two because when we did the upgrade from two x to 3, nobody cleaned up after that and the old one was still laying around. So that's fun. So I don't want that to happen again. So when we get to the final version I think the two steps will be you'll need to change the runtime. So it has the right version of Ruby and then change the bundle install step. So it's installing from the right gem file and Gem lock file.

Robert Raposa: and that's something you imagine is in the pipeline that It's not in the reborn.

Feanil Patel: Yeah.

Robert Raposa: So Dev step will also need some. change

Feanil Patel: Yeah, yes. Yes probably. But yeah, so I think you'll need to update probably your Dev stack Fork when you do that and you'll need to update configuration. I think it was where the other step is.

Robert Raposa: okay.

Feanil Patel: but yeah, I said, I think now that I understand it. It might not be too hard to get from Ruby 3:1 to Ruby three three or three four, which would improve our sort of upgrade Runway pretty significantly. And not have to think about Ruby upgrades hopefully long enough for us to actually fund replacing this with a python implementation or something which has been a pipe dream for a long time, but maybe we can make it come true.

Feanil Patel: Yeah, I'm gonna focus on seeing if I can get the mango version upgraded first because I know that's an actual. Issue that we'll run into it the sumac.

Feanil Patel: and that will require more infrastructure changes on your side because I think that makes things a little more complicated because you'll need to upgrade your deployed Mongo instances, which I don't know how they're managed these days.

Feanil Patel: Cool, and that is everything that I have? to people have other things

Feanil Patel: I'll go ahead Pats me. So did I pronounce that it? All right?

Fateme Khodayari: Problem that I'm going to mention is related to this working group or not. But I have this issue with some of the nfcs. It's actually two issues. One of them is that they don't support right to left for some languages like Arabic version and languages that must be written in RTL formats. I guess discussion still has this problem and another problem for me is that since the learner dashboard them if he has been released and

Fateme Khodayari: it's the default dashboard for I guess since quiz was released and It doesn't support a lot of languages. One of them was Persian and while ago I tried to add Persian to this make a front end, but it just gotten away with the changes that people had in trans effects.

Fateme Khodayari: they disabled the edx flat form project on trans effects and switch to a text translations, I guess. And in the alternate texture translations project and trans effects. We still don't have Persian as a default line as one of the available languages and someone has requested it but it still is not there.

Fateme Khodayari: in my PRS one of the guys, I don't remember his name. He mentioned that we will add Persian in the past week and it still hasn't happened.

Feanil Patel: So for a lot of the translations and language related information, the translations working group is probably a much better place to ask and coordinate work. Unfortunately, I don't know if there are others from that who are also on that working group who are here who want to chime in but that working group? Controls trans effects and is in charge of making those changes and so that is probably the better place to ask these questions.

00:15:00

Adolfo Brandes: I can confirm that it is also better to ask these there then in the front end.

Adolfo Brandes: it's more specific to their mandate. So yeah translations working.

Fateme Khodayari: Thank you. And another issue that I have is with the oras and I don't know what is the proper place for measuring them?

Feanil Patel: what's the issue?

Fateme Khodayari: One of them is about the UI and ux of the rais that it's a lot clicky and it's not user friendly at all and a while ago. I saw a repository he felt to me there's a plan to change the oras into a microphone I guess but I don't know when this is going to happen and In that repository. I saw an issue or maybe an PR. That some people were getting the work done to do every design of the user interface.

Adolfo Brandes: I also have that question.

Feanil Patel: I and Jeremy this might be a thing that I would love to get. More in for I think the information currently resides internally to you, which is the plans around the aura MFE and the design documents related to the new design for that. So maybe Spencer or Jeremy are the right people here to I don't expect you guys to have the answers on the spot, but maybe you guys could go find those answers and make those publicly available because I think that information is not super available at the moment.

Jeremy Ristau: you mean the future of the or product is…

Feanil Patel: Yeah, the future the aura product and…

Jeremy Ristau: what you're

Feanil Patel: also what's been done so far on the MFE and how much of that is live and how much of that is. future

Jeremy Ristau: So those are very http://edx.org specific questions. that question mark

Feanil Patel: the part that I care about is there's a new MFE. Is it gonna become the default are we going to change the code in the x-block to make it the new default? What's the timeline for that for opened X?

Jeremy Ristau: Yeah, when you say for open edx, not on for http://edx.org . Yeah, really?

Feanil Patel: Correct, correct. Yeah, yeah for everybody else and also the impact on the old UI as a result of the sort of investment in the new UI

Feanil Patel: I'm gonna as this is Spencer or you Jeremy the right person ign sort of to dig into this Jeremy. alright

Jeremy Ristau: Yep.

Feanil Patel: Yeah, so as you can see, I think that's information that we don't have generally but I think it's worth collecting and making public as a part of the opened X product roadmap.

Fateme Khodayari: Okay. Thank you.

Feanil Patel: Okay.

Feanil Patel: and yet Jeremy mentioned this Tim McCormick posted in the WG maintenance Channel and on discourse, but I'm just going to reiterate it When we moved the LX platform repository from the edx org to the open edx, org, GitHub automatically made redirects for git clones for the web which made it very convenient and didn't break a lot of things when we did that move. However since then,

Feanil Patel: A new LX platformrepo got created in the edx org, which means that those redirects have no longer work. So if you have deployments that are old or if you have sort of been continuing to rely on the fact that this does automatically redirect historically you should update your configuration so that you are cloning from the opennetics org directly rather than through this redirect.

Feanil Patel: I'll post the discourse link. in the meeting notes Kyle gofort

Kyle McCormick: If you run a tutor version that's older than maple. Then there's a Regis posted a little command in there that you can run to make sure that your pre Maple installation points at the open edx version of the repositories instead of the broken edx version of the Repository.

00:20:00

Feanil Patel: yeah, thank you. And that's in that same post so you can find it there.

Feanil Patel: is it Maple and before or anything older than maple?

Kyle McCormick: I think it's anything older than maple, but if you're on Maple, maybe just check anyway.

Feanil Patel: Cool great. Anybody have any other questions?

Feanil Patel: All right. Reminder was I think we're still looking for?

Feanil Patel: a couple of the front-end apps still don't have catalog owners, but I think some of them are to you interest ones that I think Jeremy you're working with Kyle to sort of get the team names all cleaned up before those can all land. So I think that's fine.

Jeremy Ristau: Yeah, it's more like the two things are happening parallel and…

Feanil Patel: Yeah.

Jeremy Ristau: it'd be great if the teams were cleaned up before the files get updated.

Feanil Patel: Okay.

Jeremy Ristau: But yeah.

Feanil Patel: Okay, cool. I'm gonna assume between you and Kyle I feel like I have a lot of trust that you guys got this so I'm sure it'll happen Xavier. I did integrate to the open craft p1s into the spreadsheet as well. So you can take a look at that all in one place now and if there are any repositories that you guys want to take on. Hopefully it's pretty easy to see that there's no other interest.

Feanil Patel: right yagneshko for it

Yagnesh nayi: Hi, I want to ask one question. When I see Spencer.

Feanil Patel: Yeah.

Yagnesh nayi: I can see there are many person to you interested. So which repo I will choose because if I choose any report, is there any conflict issue? Yeah, some one another also work in this report. So I don't share which repo I will choose because there are list of many reports and as one is priority Vice. I am not sure which repo I will choose.

Feanil Patel: So I think you're not a CC on anything right? So I think starting to contribute and support any of the ones where there are no interest and starting from Priority One down is probably a great place to go so you can

Feanil Patel: it looks like the first one that sort of That build is the edx organizations repo. so I think to help establish a history of contributions and commitment the thing to do is to start looking at the issues in that For example, the python 312 upgrade in that repo and start contributing to getting that ready to be mergeable.

Feanil Patel: start there. is my thought

Yagnesh nayi: Yeah, but also a couple of days ago I message with Xavier and He suggests me to you can pick up any repo you cannot as a currently CC, but you can check lots of things like you don't have much rights, right, but you can still see issue and…

Feanil Patel: right

Yagnesh nayi: you can resolve and then you can ping us to and you can help like this.

Feanil Patel: Yeah, yeah, Absolutely. you can contribute to any of them.

Yagnesh nayi: Yeah.

Feanil Patel: I like If there are places in the system that you're more familiar with making looking at the issues in that area and making pull requests to fix those issues is a great place to start but if you're looking for a place that needs sort of Maintenance support and health then I would start with things where other people are not already looking just because we need help in those areas and there's not obvious interest from elsewhere. So either is fine, I think To sort of establish the history of contributions and commitment. I would start wherever you feel like you're strong and where you have the ability to help.

Feanil Patel: so either strategy will be fine. If you're looking for help my perspective tends to be from the maintenance perspective right now. So I'm always going to tell you to look in places where there aren't already other eyeballs because I want to make sure we get full coverage of the things that we are releasing.

Yagnesh nayi: Sure.

Feanil Patel: But anywhere is fine any help is good help.

Yagnesh nayi: Okay.

Feanil Patel: Does anybody else have anything?

00:25:00

Feanil Patel: That's my go for it.

Fateme Khodayari: I have another issue as well some while ago.

Feanil Patel: Okay.

Fateme Khodayari: I realized that the download link for Ora submission files. Which is coming from the backhand side to the front and side was incorrect. And I guess this got fixed when Korean Spencer's release, but recently I found another issue which I feel again. It's a URL sent from back into front and it's about the view grades link that is displayed in dashboard that's on one hasn't passed the course and…

Feanil Patel: Okay.

Fateme Khodayari: is not eligible for certificate. There's a button or…

Feanil Patel: Okay.

Fateme Khodayari: a link on the dashboard that it's displayed under the course card. And the problem with that is that it's redirects the user to the old progress fate.

Feanil Patel: This sounds like it's worth writing up as an issue on the edx aura GitHub repo which is in terms of reporting sort of specific issues with specific components. We track those via GitHub issues across the project. So I would suggest that it sounds like you have a really good understanding of how to both produce this problem and what the problem is. So I would suggest creating a new issue on the edx or at your repo.

Fateme Khodayari: 

Fateme Khodayari: I created a topic in the forearm some while ago,…

Feanil Patel: Okay.

Fateme Khodayari: but it got no replies. So I guess maybe there are other places as well that's back in descending a link to front end which may be wrong. So I guess maybe it's worth having a look at all the links that are sent to the content.

Feanil Patel: Yeah.

Feanil Patel: Yeah, I mean, I think that's true. I think aura

Feanil Patel: two currently doesn't have a clear maintainer.

Feanil Patel: Although it. Looks like maybe there is interest I think to you is gonna take that over it looks like So I think reporting the Jeremy correct me from but I get up issue feels like a much more actionable place. So if you already reported it, it might just be a matter of creating a GitHub issue that links to The Forum post so that we have it closer to the code. if you want to drop the link to that discussion for post here. I'm happy to make the GitHub issue on this as well. Just so we don't lose track of it.

Feanil Patel: And you feel free to send it to me via slack after the meeting also if the meetings before you find it.

Feanil Patel: All right. other topics

Feanil Patel: with everything Thank you everybody for joining us. I think we'll sort of swap over to edx platform Details next if you want to stick around for that. You're welcome to but if not, feel free to drop off.

Feanil Patel: And I'll see you guys next time. All right.

Xavier Antoviaque: See you.

Feanil Patel: 

Feanil Patel: all right, and X platform

Feanil Patel: I think the biggest thing to cover here as far as I know there's the node upgrade stuff which we covered already, but the other we just stay here Jeremy. It's

Jeremy Ristau: I know I forget everything.

Feanil Patel: Okay, the biggest thing to but I'll update that meeting. So it's just the same place. So you just log back in automatically. but yeah, I think that the big things here that we've already covered is that the note upgrade which is pretty behind.

Feanil Patel: the python upgrade which

00:30:00

Feanil Patel: the python upgrade there are tickets that awaisins team made or not. They've made full requests for Python 3 9 10 and 11 already so that we can see what the failures are in the tests in platform always is are you guys planning on digging into fixing those issues as well or was this just like do you have capacity for that or is it just like here's as much as we can do right now and somebody needs to go dig into them.

Awais Qureshi: As you right now we are focusing on site packages. So once we complete all those patterns then I can try to fix those, but right now it's just for saying that the tests are running.

Feanil Patel: but

Awais Qureshi: And I think the failures are not much but they have some major failures this cash. So once we fix those issues, I think they will be less failures.

Feanil Patel: So when you say package you mean the dependencies of edx platform? Yeah.

Awais Qureshi: This is We are working on we have already created two dozens here and…

Feanil Patel: Okay.

Awais Qureshi: we need someone to review and merge them. So yeah.

Feanil Patel: Yeah, so I think getting those Library PRS are viewed as should be top priority. So think I don't know if everybody has any capacity for any of that or can do that in between other stuff, that's great. I've asked accent improvements just start looking into those once they finish the things that are under accent maintenance and hopefully plus Community will I'll put another note in the CC's Channel about it as well to get them to start looking at some of these and big and getting Over The Line

Jeremy Ristau: Yeah, and I think the current state of the RB Bond teams is to attack and deliver all the things that are going forward owned by us. And then only…

Feanil Patel: Okay.

Jeremy Ristau: if bandwidth is available after to pick things up.

Feanil Patel: That sounds That's fine. that sounds fine. I think hopefully those PRS are good. Yeah. Yeah.

Jeremy Ristau: Yeah.

Feanil Patel: and then Any update on the mango upgrade stuff? I think that's the big thing that

Feanil Patel: I don't think I have enough info to sort of start moving forward on it and I think maybe arbaum is the right people to ask in terms of getting it tested on newer versions.

Feanil Patel: So yeah. I don't know if Robert or Jeremy or Spencer like the Mongo upgrade. I think we just need the tests to start running on the new versions before we can dig any further and I don't know what the blocker is to sort of have that happen other than capacity, which I know is the perennial blocker but

Feanil Patel: Is it go for it?

Robert Raposa: I'm just gonna say I'm not aware that maybe there was Jeremy can speak if Jeremy's Ready, but I'm just not aware of arabicum has that on their list of to Do's or…

Feanil Patel: right, so Yeah.

Robert Raposa: if testing for mango, I just good.

Feanil Patel: I think last time they said there was a Docker file that needed to be updated to have the new Manga version on it. And I don't know which Docker file that is. And if I knew which one I would just make that pull request.

Robert Raposa: So I think since we last met I don't remember one that's happened. But then as I really landed the Mongo upgrade that affected Dev stack and so I don't know if that's what needed to land and has landed.

Feanil Patel: right

Robert Raposa: or if that's something different but that upgrade landed and broke everyone's steps back until they get the instructions of how to upgrade and

Feanil Patel: Okay.

Robert Raposa: and I don't think that was a death. I don't remember if that was a Dev stack specific PR. That was the pier just

Jeremy Ristau: Yes, They're going environment by environment. So that was Dev stuff specific and then they did stage. day before yesterday

Feanil Patel: Do you know when they'll do CI because that seems like that should happen before Dev stack.

Jeremy Ristau: Okay, find out today. just to be clear though, Most of the work for Mongo upgrade is being done by the SRE team not by the bomb teams.

Feanil Patel: God I got it.

Feanil Patel: Okay, let's get to know.

Jeremy Ristau: We're just staying informed more than anything.

Feanil Patel: God I got it.

Robert Raposa: and does that just I affects. What repose? it just

00:35:00

Feanil Patel: Just at X platform as far as I'm aware. And it didn't need to…

Robert Raposa: Okay.

Feanil Patel: but the way that it was previously implemented means it does long term. I want to change it so that there are GitHub act reusable workflows and GitHub actions to install whatever version of Mongo you want as a part of the GitHub action and I assume we didn't use it because we thought it would be faster to have it pre-installed, but I've tested some of these workflows and it's pretty fast to get mango installed at whatever version so moving forward I think I would love to move to that and have The runners be more vanilla, but that's a bigger change that then I can like effect in the short term.

Robert Raposa: so if this is just an issue of a single repo. Is this something that's in the GitHub actions that anyone could do or no?

Feanil Patel: No that's the problem. Is that in The workflow runs on an image that is made like a Docker container that is maintained by to you currently. And so we need that workflow the container that the GitHub action Runners are the private Runners for edx platform and we need those to be updated right now.

Robert Raposa: yeah, and Jeremy are you clear if that is on sres list and we could just move it to the higher up on their list or if that's

Jeremy Ristau: I would recommend talking to Farhan numer. I don't know what the SRE status of Mongo greatest.

Jeremy Ristau: Or like that. I don't know the tasks.

Feanil Patel: I'll have to Administration.

Jeremy Ristau: They have listed in their back one for it are already.

Feanil Patel: And cool,…

Kyle McCormick: I think correct me…

Feanil Patel: I will.

Kyle McCormick: if you disagree for Neil, and if another resolution of this is if you folks can Open or find that in the edx org and transfer that repo or that darker file or whatever. That's building that private image to the open edx org, then it's something that the community can handle.

Feanil Patel: write it sounds like the people we need to be talking to is Farah numer the right person on SRE it to be talking about this. so it sounds like it's not in the hands of anybody who's attending here and if I can't find them in the open X slack I may need for you to connect us. But if I find them I can talk to them directly.

Feanil Patel: Cool, I think that's great that at least helps. Figure out what we need to do to move forward on that.

Feanil Patel: Any other major edx platform ownership things that we want to cover today?

Robert Raposa: one topic that I forgot about That is not edx platform specific necessarily, but do we want to discuss but it's more to you specific. Maybe do we want to discuss the select thread To changing of admin unrepose and timing of that and implications of that were.

Feanil Patel: Sure, what do you guys want to talk about? Yes, so context for everybody's on the same page is historically various to you teams had administrative access to specific repos. And when we moved those repos to the open edx, or we didn't go update everything just because we didn't have capacity to do it with all the more recent changes and sort of to get in alignment. This is the thing we needed to do eventually It eventually just happens to be now. So we're sort of moving forward with that change. The plan is to convert all of those admin permissions to write permissions. So any day-to-day flow should still work.

Feanil Patel: But if are there specific concerns with that or what are the communication channels?

Robert Raposa: my yeah, and I'm specific concern around timing of Eventually is now. there's going to continue to be I'm guessing. things Decisions that affect to you and you have to do a bunch of work too to handle them. and…

Feanil Patel: Yeah.

Robert Raposa: I think that's a reality and that's fine. What I'm wondering is when we discover that this is something that's going to have us having to dance around and do a bunch of stuff. Is that something that we can better plan together so that it's not like, we're doing this thing today and now go deal with it because right now as I said,

00:40:00

Feanil Patel: Yeah, and I'm sorry if this wasn't clear in the original communication, but I don't think we'll find what those things are until we do it. So we're gonna do it and we can just move it backwards if it's a critical enough issue and the time and you guys don't have capacity to fix it right away as long as we can then commit to a time when it will get fixed. we're in this ambiguous eventually and what I want to get to is let's sequence all the work that we're aware of so that it can get done in a timely fashion, but personally I don't have everything needs to happen before March 1st or everything needs to happen immediately as long as we can talk about it and get it all done. So I'm happy to be like, hey, it turns out that this admin on this one repo is super required and we don't have a good answer for how to work around it in a short term.

Feanil Patel: I'm happy to put it back. We will just collect that information in a ticket so that we know why we did it.

Robert Raposa: That sounds good. where should we have both the specific conversation around how we're going to Resolve short-term issues around this and…

Feanil Patel: Yeah.

Robert Raposa: then also the Lord more General here are the things that need to be sequenced and planned.

Feanil Patel: He yeah, I mean that's a great question for this specific issues. they should come in as accent requests. And that way we can track them along with any other permission changes that we do for everything else.

Jeremy Ristau: So I think this was the concern that Robert was bringing up is at the time at which we discovered that we Elevated access it is like the not a three day file an accident again, wait a situation.

Feanil Patel: Yeah, yeah. I'm happy for you guys to file the ticket and immediately escalate to me. For these issues, but I want to get them tracked as tickets so that we can find them later.

Jeremy Ristau: Yep.

Feanil Patel: So yeah for these specifically Feel free to escalate directly to me and if I don't answer, I'm gonna throw Sarina under the bus and escalate to her. but I think that we want to track them all as action requests tickets if there are emergencies. I'm happy to give you my cell phone number.

Feanil Patel: And then Robert already has it. So

Robert Raposa: yeah.

Jeremy Ristau: And is there so they own this workstream popped up as an email? Not as roadmap item or…

Feanil Patel: yeah.

Jeremy Ristau: anything. is there an issue that it points back to?

Feanil Patel: Yeah.

Feanil Patel: so Before the big shake up. there's this edx side Open Source process working group.

Jeremy Ristau: Yep.

Feanil Patel: Repo and project and that project was where we were tracking sort of What these items were to you and ask them in particular need to work together and sort of sequence and then everybody who was in that meeting? Doesn't exist anymore. So

Feanil Patel: yeah, you're but that was where we were tracking that so we can start using it again. I think the goal was that you guys would come back to us sort of if that wasn't where you wanted. It tracked or an alternate proposal or how we want to revamp this meeting. But I suggest that that particular backlog get cleaned up and then I'm happy to make these things as new issues on that backlog so that they don't come in Via emails or slacks or all these other places, but I think we had sort of. Dropped that for the time being because you guys are figuring out sort of what you wanted to do internally. So I suggest we've sort of revive it and clean it up or I'm happy to make a new GitHub project on the open edx side that I can easily manage because this is a thing you guys gonna have to manage.

Feanil Patel: Whatever is I am open to process suggestions on this but I don't want to wait for those processes to exist to start making movements.

Robert Raposa: and do we need a separate meeting for dealing with that or it can

Feanil Patel: Yeah, absolutely. Yeah, absolutely because that's gonna be broader than a next platform and it's going to be broader than what the red of the rest of the community sort of needs to like. I'm okay with them following it, but I don't think they need to be aware of it, especially because some of it is going to be cleaning up differences between permissions that we've had forever and what we want to move to and they're going to be specific to you. So

00:45:00

Feanil Patel: and I'm sure there's going to be a messy Middle where something is actually for everybody and we had a conversation in the one meeting or this where it's too you specific but we're having the conversation in the edx platform meeting like that's gonna happen. I'm okay with that, but I think we should have a separate space to talk about to you and to you specific things that Are these transitionary things? And I would love for that to be a meeting where our goal is to eventually not have that meeting because you guys have the same sort of cover like you guys are in the same breath as the rest of the community in terms of how we work.

Robert Raposa: Yeah, so it's just figuring out if we need a new meeting with the same people or if we can make the group then edx platform specific maintenance and then, if there's time left over we'll good.

Feanil Patel: Yeah. Yeah, I'm sure I think as long as we add a bathroom break. I'm sure we can tack on. That conversation in this one. I think the maintenance working group meeting. Will probably get shorter. it didn't take the full hour today and I think that'll sort of remain true as we get a little bit more efficient with it. We may not need it all the time. So my goal is to eventually not have that every week. But right now there's enough things in flight and enough information that needs to flow that is the fastest way to get it. But yeah, I'm happy to have that be sort of like track three of this meeting which is and now any edx to you specific things. We want to cover. And I'd prefer to still record it and that way if people want to follow along they can.

Robert Raposa: so on the other thread If I want to brainstorm ideas. are you saying I should just create a public engineering ticket and frame starting the ideas on because I want to be able to have the kind of a discussion can we have a temporary solution where there's a specific admin group that might not have everyone across, just has a couple of people in it that are in you that are sort of trusted and…

Feanil Patel: Yeah.

Robert Raposa: that are And escalation path for people, you know what I mean…

Feanil Patel: Yeah.

Robert Raposa: if you want to serve sooner rather than later, but where do you want to discuss that in a meeting like this? Do you want to discuss it on the ticket lack? just as well that discussion before a trigger is pulled.

Feanil Patel: I

Feanil Patel: right Are you asking about this specific issue or you asking about where should we do this generally?

Robert Raposa: This specific issue now.

Feanil Patel: for this specific issue we can just talk about it right now

Robert Raposa: How do you feel about that proposal?

Feanil Patel: I think it increases the amount of work we have to do and we don't know if it's necessary yet because we haven't done the cleanup and I think if we find an issue and immediately return admin access that feels like fast enough to me that we don't need to create a new team at people on that team train that team only to have a plan to then remove that team in the future. But if you guys think that there's enough risk on your side with going from admin to write that that is necessary. I'm happy to do that, but it feels like extra Hoops to jump through from my perspective.

Robert Raposa: Yeah, I mean I'm just concerned with the people who had mentioned. this is a situation that has come up in the past and we did need Emergency access to Xavier Z. And now we know we won't have that access but we won't know about it until said emergency happens in the future and so just feels like We need something which might be not doing anything until we have this future plan doing some. change with a different set of admins or

Feanil Patel: Yeah, I mean again I think. if you have specific examples of this repo needed this thing in this emergency, I'm happy to sort of discuss that but one of the hard Parts about the situation is that we don't remember what those historic emergencies were and we don't know if they're relevant anymore. And so that feels like I'm not disagreeing with you that there's risk. I just think it's so unknown that it's hard to plan for in this situation and we can always make it. we can do more work, but I don't know if it's necessary. So I'd rather not.

Jeremy Ristau: I think for this there are only two I totally agree with you. There's a little bit too much risk aversion with this conversation right now, I think and ultimately what we need to do is move forward break something fix it as fast as possible and learn from it.

00:50:00

Feanil Patel: Yeah. Yeah.

Jeremy Ristau: I have two questions associated with this kind of risky change one can you get me an exact date where this is going to happen instead of in the next few days kind of thing. That way we can point teams to say before the state after the state.

Feanil Patel: Absolutely.

Jeremy Ristau: If you see changes in Access, to escalate immediately to finale and…

Feanil Patel: Yeah.

Jeremy Ristau: that would be a 24 hours a day situation escalate to finale.

Feanil Patel: Yeah.

Jeremy Ristau: Yeah, and then my second question is more around risk mitigation versus End goal here I do. Let me State this as a question. Do you see an end state where One by one re request admin access for every single GitHub team and then you granted one by one and then we're in the exact same state that we are now, I assume that's not the goal.

Feanil Patel: no, yeah.

Jeremy Ristau: So what you're offering is a temporary workaround in order to deal with an emergency. I don't necessarily Think that is a long-term solution.

Feanil Patel: Yeah.

Jeremy Ristau: So what I would look for is what is the long-term solution here? That I think I need in hand when I go and I tell people about this. I need to say you have a temporary workaround if you run into an emergency, but that is not going to stay forever for whenever an emergency appears. What is the real long term solution?

Feanil Patel: Yeah. Yeah.

Feanil Patel: For the first thing yes, I can give you an exact date. I can make all of those challenges happen at once.

Jeremy Ristau: Yeah, I think.

Feanil Patel: I'm happy to be sort of 24 hours on call for some specific defined time after that date, which is not forever. And then I'm gonna do not disturb on when I go to bed after that, but For the second thing. I think that is the thing that is hard to do without knowing what it is that the administrative access is used for and why it is necessary. Because those Solutions are going to be based on what the problems actually are in those spaces.

Feanil Patel: Some of the known issues that I can imagine are we need to be able to quickly make changes to Branch protection and answer for that is unfortunately you cannot do that as quickly as you used to be able to because we need to be able to track those in a centralized manner. They have to go through us like that. That is just gonna get slower other things might be we want to be able to add outside collaborators quickly so that we can get things landed and…

Jeremy Ristau: Yeah, yeah.

Feanil Patel: that's like a you're not supposed to do that. And so we didn't want you to do that. So we're never going to let you do that again.

Jeremy Ristau: I wouldn't work for us either from a soft complaints here.

Feanil Patel: Yeah. Yeah, I figured but the examples are sometimes it's gonna be like this is slower. Sometimes you shouldn't have been doing this in the first place and with the third thing where it's like, okay, we make sense for you to be able to do this quickly. I think we will need to develop tooling that lets you do those things without doing the rest of the things or we may need to develop automation that will let us ensure that that's true across everything so that you don't have to go in and manually make certain changes. So I think identifying what those things are that are these are an example of this would be like the checkbox that automatically deletes a branch when it merges to main or master.

Feanil Patel: It's in most of our repos, but it's not on and all of them and it's nice to be able to go in and turn that on when it's off. But the thing that's actually the solution is that should just be a default across the entire organization that is applied and true, but it's not right now and there might be other things in that vein of things that people have been turning on in certain repos, but maybe you should just be a discussion and turned on everywhere, allowing Auto merge as a feature is I think one of these things that is in a handful of repos, but it's not on everywhere and that should probably be a discussion rather than somebody going in and turning it on just because

Jeremy Ristau: Okay,…

Feanil Patel: 

Jeremy Ristau: so I would if I were to say back to you what I think I hear. this is the first phase of learning to discover if and when admin access is needed.

Feanil Patel: yeah.

Jeremy Ristau: And admin access will be granted when needed. So that we can learn from that situation but there is additional work that axim or the community needs to do probably action here specifically needs to do in order to have an actual long-term solution for this. It okay.

Feanil Patel: Yeah. and with

Jeremy Ristau: All right, because the communication definitely read this is the long term solution. We're ripping everything up. So I just wanted to be yeah.

00:55:00

Feanil Patel: Yeah, yeah, and to be clear the long look like the long term solution may be less convenient. Yeah, …

Jeremy Ristau: Yeah. totally. I don't know all about that.

Feanil Patel: but yes the long yeah.

Jeremy Ristau: I'm all about that. Yeah.

Feanil Patel: yeah, it may not be as good as it was in terms of what you think is the fastest way to do things but it is what we think makes the most sense for the project and yeah I want to work with you guys to get to that but I didn't like there's no place to start except by doing it. I think.

Jeremy Ristau: Yeah makes sense.

Feanil Patel: All right. Anything else you guys want to cover?

Kyle McCormick: Sorry small thing away. So you still there?

Kyle McCormick: Nope, he's gone.

Feanil Patel: I think he dropped up yet.

Kyle McCormick: Never mind.

Feanil Patel: All right, And that exact date will probably be sometime next week. Once I've finished scripting up the thing that will update all of those admin rates. But I'll tell you the exact day when I have it. Yeah.

Jeremy Ristau: Possible could you and that as early in your days as possible?

Feanil Patel: Yeah.

Jeremy Ristau: Just so we have one whole working day to get the blockers off. Yeah.

Feanil Patel: Yeah, I'm very flexible on when it actually happened as long as it happens.

Feanil Patel: All right, anything else you guys want to cover?

Robert Raposa: but Spencer just that there was a topic that we were talking about you bringing here, but I don't remember what it was or if we covered it.

Spencer Tiberi: Don't either. Sorry, Robert.

Feanil Patel: Yeah.

Robert Raposa: For giving one second. I'm just gonna look at my history

Spencer Tiberi: Sounds familiar, though.

Feanil Patel: I mean the last thing Spencer that you posted in maintenance group was that just upgrade stuff. Is there anything specific to that that you wanted to mention?

Spencer Tiberi: No, I don't think so. I'm just helping communicate. That it's all yeah.

Feanil Patel: Okay, great.

Feanil Patel: All Okay, I think I'm gonna let you guys go then I think if there's anything else just drop it in this channel.

Robert Raposa: Yeah,…

Feanil Patel: Robert.

Robert Raposa: I think it was so Spencer wrote.

Robert Raposa: So PR came up related to tutor. and…

Feanil Patel: Yeah.

Robert Raposa: he said, will we on these bugs long term and there's anything Tutor related become community-led. So it's related to the questions that we had earlier around getting tutor testing on everything and how that works.

Feanil Patel: Yeah.

Robert Raposa: And what happens with Tudor bug tutor related bugs.

Feanil Patel: So tutor is the community supported release…

Spencer Tiberi: Thanks, Robert.

Feanil Patel: what that means is if you are maintainer of a system in the community, you're responsible for making sure your component Works in tutor. So if tutor related bugs come in those are bugs that you have to deal with as a part of tain maintenance of the components that you're maintaining.

Feanil Patel: That answer your question.

Kyle McCormick: I think

Robert Raposa: I mean, I think it answers it in the gym in the general now. It's like I've never used tutor so I know some others have for dubstep whatever but I don't actually know what that's going to mean for us. So, I don't know if that's simple enough.

Feanil Patel: Yeah, I don't expect it to be simple, but I think That is a part of what you're signing up when you're maintaining a thing is that it works in tutor?

Kyle McCormick: I can add a little bit of color there are times where they're tutor plugins or tutor itself will have bugs. And in those case the er of the Plugin the maintenance of tutor are on the hook. But there are times when the repo itself is doing something that makes incompatible with tutor. and I think ideally there's a collaboration between tutor maintenance and tutor plugin maintenance and application repository maintenance To determine what is a bug in which system and there are times where it's both? Yeah.

Robert Raposa: Yeah, so can any be brought to the tutor maintenance and I guess that's what you said is on the forums just to even find out to help triage that in some way of where we think the bug lives.

Kyle McCormick: Yeah, I'm making this up on the fly right now. but issues in repositories are a good place to track bugs we can make it in one repo and then transfer it to another. If it turns out the issues in the wrong place. I think there's a page that lists all the tutor maintenance and…

01:00:00

Feanil Patel: right

Kyle McCormick: tutor plugin maintenance that all send along I'll put in the notes so that can be a good way for application repository developers to connect with those two to maintenance.

Kyle McCormick: And I think in the other direction of two people using tutor find issues, they might file them on the forums or they might file them in the tutor repos or my file them in the open edx Repose. So

Robert Raposa: Okay.

Feanil Patel: I think there's always going to be a little bit of triaging and moving the bug around to find the right responsible home. and I'm to read back. I think it is. Okay as the maintainer of a repository to be like this is actually a problem in the plugin and not in the repo and should be sort of the responsibility of that other person and that conversation should then be had with the maintainer of the plug-in. But for example with the Ruby stuff when I land this change, I'm going to go notify the tutor maintainers that the version of Ruby that the forums runs on should be updated to this new version and then I expect them to handle the plug inside of the work from that message.

Kyle McCormick: similar to the relationship between

Jeremy Ristau: but in this particular up

Kyle McCormick: Was that similar to the relationship between application developers and Dev stack developers in the past where if debtstack broke maybe it's Dev stack. Maybe That's the problem. They're probably a conversation between the maintenance or the owners of both of those to determine. Who should do the work or who needs to do the work? That's just shifting now between repository developers and tutor maintenance

Jeremy Ristau: Question and…

Feanil Patel: the Jeremy or simply

Jeremy Ristau: yeah around your arm. the example for Ruby, you as the person doing the upgrade is letting all of the stakeholders know in disparate channels that this is happening is the expectation that any worker doing this kind of thing would have to and know how to communicate.

Feanil Patel: This is ideally a thing to maintainer would do but there is no name maintainer for CS comment service right now. you is ostensibly interested in Kelly's team. I think to maintain CS comment service but they've not been participating. So I've just been ating. Without them at the moment, I think longer term I'll engage them and make sure that this is a thing that's documented this particular thing. I think there's not a lot of stakeholders, but be the two sort of big ones that I'm more of are essentially tutor. And to you, right? and…

Jeremy Ristau: I…

Feanil Patel: so those are the only

Jeremy Ristau: This is a general conversation with a specific example, but it sounds now more like General contributor contributes maintainer must see that contribution and…

Feanil Patel: Yeah, yeah.

Jeremy Ristau: then know who to communicate out that contribution to that should at least added to the maintenance of things.

Feanil Patel: For major updates yet. Yeah. Yeah, that's a fair point. Yeah. I think this is very specific to Major upgrades and we're sort of still figuring out what the right. Task is and I'm kind of making it up as I go.

Jeremy Ristau: Kind of I mean, hopefully except when we want to remove the user ID from the UI or we want to do other things that have really big impacts a maintainer other repo may say yeah.

Feanil Patel: Yeah.

Jeremy Ristau: I'll just let it happen and not tell anyone. That's a stakeholder and so all the process kind of happens,…

Feanil Patel: Yeah.

Jeremy Ristau: but ultimately nobody. the stakeholders didn't approve.

Feanil Patel: right …

Jeremy Ristau: You know what I mean?

Feanil Patel: right so for a lot of the things that are sort of more in that vein, that's where the product working group actually comes into play and longer term. We want to move more of these things so that they're tracked under that umbrella and that process so that before things disappear off of user interfaces. There should be a product review and then product is in charge of doing all the comms externally to them and sort of tracking the development behind them. They become that interface. so that you don't as the engineer have to know to notify everybody about everything.

Jeremy Ristau: Okay, Peanut gallery Comet it sounds like it's getting slightly complex. We're like…

01:05:00

Feanil Patel: Yeah, no, absolutely.

Jeremy Ristau: if it's an upgrade Engineers communicate, but only container engineers and if it's a feature change product working group needs to know…

Feanil Patel: Yeah.

Jeremy Ristau: who to communicate to it does sound like it's getting a little bit. complex

Feanil Patel: Yeah, no, I absolutely agree part of the conversation. I'm having with the product working group is how and we've actually started doing this on the product roadmap is we are starting to track the major upgrades along with the product stuff so that it's getting communicated through that same process. and part of I think what I want the maintenance working group to do is be this channel where we can tell them these things need to happen. And also then when those things happen figure out what those Communications channel are and be the clearing house for a lot of that. So that individual Engineers don't necessarily need to keep track of the 17 different places where somebody might care about this.

Jeremy Ristau: yeah, because I'm afraid I need to go back and…

Feanil Patel: community

Jeremy Ristau: start watching more discourse subsections that are product related instead of engineering related because maybe they're posting things there or maybe they're posting them at slack and me …

Feanil Patel: Yeah. Yeah.

Jeremy Ristau: just it's spiders out so

Feanil Patel: No, absolutely. I think my goal is eventually that you can watch the product roadmap and be aware of all of the things that are happening whether they are technical architectural changes or user facing implementation changes. That is like the North Star. I'm moving towards I will be messy but between here and there, I'm sure.

Jeremy Ristau: Okay. Thanks.

Feanil Patel: Yeah, all right.

Feanil Patel: And leave a minute for anybody else or anything. I think we're good. All right. Thanks all this was super useful appreciate all of the feedback and pressure testing of all of this.

Robert Raposa: Thank you so much.

Feanil Patel: Yeah. All…

Jeremy Ristau: Same absolutely. Yeah.

Kyle McCormick: Thanks, everyone.

Chintan Joshi: Good. Thanks.

Feanil Patel: Have a good day.

Chintan Joshi: See you guys.

Robert Raposa: again

Feanil Patel: but

Meeting ended after 01:07:19 👋