Open Telemetry Uncategorized, 22 Sep 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2022-09-22 meeting

Description

cncf-opentelemetry@cncf.io's Personal Meeting Room

A

A

No hello, everybody looks like Philip and Severin are running a bit late, so.

B

B

A rough time here.

C

A

C

Some just give me a second to settle.

C

D

C

I just wanted to say: let's get to live a minute or so to join, but there he is hello. Yes,.

E

Sorry for being a little late, different meeting ran over a bit yeah.

D

E

Customer data always fun.

C

I I had to fill the washing machine and it took me two minutes longer than planned, so not that urgent, awesome, let's get started I think we have a bunch of things on the agenda already. uh uh Phillip I think the first point came up from you, so maybe you want to take it from there. Should we put it on. Let me traffic and share my screen now. It's uh soon.

D

E

Awesome right, yeah I, should probably fill in that placeholder link that um isn't really a placeholder anymore.

D

E

D

E

um Yeah this this was a discussion topic in the maintainers meeting last week.

E

um They do like the whole Sig check-in sort of thing. I introduce myself to folks. I, also introduce you um uh that you're you're, also one of the maintainers and uh the major thing is going to be metric stocks because they don't exist and in fact we uh like pulled unmetric stocks that were wildly out of date and Incorrect and leading people down the wrong path, uh and so I'm like hey.

E

uh We, we should probably have some docs like maybe before kubecon and they agreed uh and what I sort of told them was I'm, like hey, really kind of the the thing that's hard with otel. If you don't know how to do, something is usually initialization because, like once, you have stuff initialized correctly or like in the most correct way.

E

You can kind of just sort of follow, along with the API and like it's usually not too hard, certainly not from from my perspective, to like add docs for things um but like using golang as an example. It is, it was way harder to learn how to initialize open Telemetry tracing than it was to like actually like hey. This is how you create a Spam. This is how you create a span event. This is the greatest band link like that was like 10 times easier and really easy to do um so.

E

I told them that's sort of the dynamic here so, like you know, um I'm willing to accept lots of contributions from language sigs to like have um metric stocks, but if they're like well, you know. Here's like a dump of like a bunch of code and like this has examples of like all the ways you want to do it.

E

We don't know how this is going to look best in docs, but like this is kind of like all the major stuff like we can take that and we can morph it into documentation pretty quickly and so um uh Tyler on the gosig and Dylan on the JavaScript Sig both said that actually sounds fantastic. They would love to work in that model and so they're just like.

E

If you could give me like a high level, what the overall intended structure is like what just like the major Concepts you want to have documented are um with just like names like we can fill in some blanks with some copy, pasteable code samples and then like hand it off to you and like we can iterate on making it good.

E

So um that's what this is I created in uh an issue to ask that you had uh lay down your opinion on kind of how to generally structure things, and so I just put some comments in there of like okay.

E

You know these would be like the headings for the manual instrument right, yeah and like the getting started page and that kind of stuff, and so like, for example, on the getting started, one there's more than just metrics on there, but like that, that's like where the metrics part would fit into a getting started page for a given language um and then, if you scroll down to the the other one, this would be like for manual instrumentation. There's like a high level section about tracing.

E

That's like pretty much all filled in for most of the major languages today, so I didn't bother putting stuff in there but like with metrics.

E

This is like where we would need some help um and, like bullet pointed these, are the sorts of things that we need to have in a code sample that will like set people off in the correct direction in initializing metrics for the first time and then like from there, it's pretty easy to fill in more of those blanks. um If they have code samples, that's great I'll. Take them if they don't have one for like using an upbound counter. I could probably create one in five minutes uh and yeah.

E

So I, don't know what do you all think.

B

um How am I my recollection is: there's very little overlap between Trace initialization and metrics initialization.

B

E

That might be accurate for some languages, I I can't I can't say for sure like in Python. It looks pretty damn straightforward, uh but that's just python.

B

Right: okay, sorry, let me rephrase that.

B

There's no shared Logic for like like, for example, if you were doing app metrics from python, you have to configure otlp export for both traces and metrics independently, with like some notable caveats here. If you're doing all in bar config.

E

Yeah and that's also the case for for net as well like you, you make a call to like add open, Telemetry tracing, and then you have like another three metrics.

B

Yeah so I know by the way, no like, like I, think this is good in terms of like how to build it out. I'm wondering if from a presentation standpoint should we uh should y'all ask sorry would it be make sense to have like.

B

Some sort of separate tabs or like independent sections on the same page.

B

So instead of like one long thing where it's like tracing the metrics and then eventually logs, it's like tracing tab, metrics, tab, logging, tab.

B

So you've got a python, slash, quick start and then there'd be tracing. And then, if you click tracing, you would be in the tracing context and if you click metrics you'd be in the metrics context.

E

uh We could also that could potentially cause a problem for both Python and JavaScript, um more JavaScript, maybe less python, but we want to add typescript stuff to everything, and so that would likely surface as tabs. And so then you'd have oh.

B

I see like nested tab, stuff yeah that would suck well I guess we have I guess several Pages was fine. I.

C

I'm just wondering what what so so I mean, even even now with metrics, but but it's nevertheless, the usual flow that people say like hey, I want tracing and then I want metrics, so I don't know where logs come in, but to me at least it feels natural that they say like okay, let me set up tracing. Let me play around with tracing okay. Now I have tracing in place. Okay, let me add metrics in in that order.

B

I mean I think that it's one of those things where we're just we have a bias, because that's the order hotel is doing them in.

C

Yeah, but it's also the order I find the most valuable for for people, I mean they can have metrics also with.

B

I agree I'm just saying.

B

I think part of it. So speaking as someone that's gone through a lot of tutorials recently.

B

When there's you know, I think the risk in my mind is that if we have all this.

B

On like one page, one giant quick start page, then it just gets really hard to parse.

B

And so we need to have it split out. Oh.

C

I I mean it's a fair point, I'm just wondering if this is really the case. For all the pages I mean I can only also speak for for JavaScript and maybe Java I. Think setting up uh tracing and metrics is not a lot of work right.

C

I mean you just say like hey here's, my my tracing endpoint, here's, my metrics and point done and then even some of the libraries just give you that out of the box, maybe for manual instrumentation, maybe that we should really do SUB pages and say like hey: here's how you do tracing and then because it's a lot more text right, I mean um but but I think for the getting started.

C

Aka doing things automatically, maybe there's easier to say like okay, here's how you get all of it, everything yeah, um but but maybe we can evolve from there, I mean. Maybe we can take one of the languages see how it looks like in in one Big Blob and then, when we have the feeling like hey this, this doesn't look good yeah.

C

um Then then, we can still say like okay, let's re-slice it and take the metrics out into into an individual page. Okay, um I think what I? What I like about the structure that Philip laid out here is really that um I I think I put a comment there. I mean this.

C

This is this is more than metrics I would love having every every language looking the same right I mean we discussed this in the past as well, as we said like hey, wouldn't it be great if they would have a very similar structure like the example is always an HTTP server. It's always like here's. How you do automatic here is how you add uh a metric for this dice or whatever uh thing Titan had, or maybe maybe another good example, but but but giving people something repeatable and uh relatable.

C

So yeah, that's definitely what independent of metrics! What I like about this structure.

B

Yeah, no I think the structure is really solid.

C

E

Okay, um I'm thinking just because uh um Dylan and Tyler like talked about wanting to contribute stuff like we'll just start with them, and to go in JavaScript and kind of see how it goes. And if we find it's like too wild um I guess we'll figure something out.

C

Okay, yeah, no, no I think that that's probably a good idea just to to pick one or two languages and then um I'm, just wondering I mean JavaScript with python I mean both are I'm, always wondering. Maybe we should also have something like go or rust. I mean their documentation is non-existent, uh but but something that that's a little bit more like a compiled language, where we don't have that much around automatic instrumentation, but yeah I, don't I, don't know I mean if we have Python and JavaScript already wanting to get started.

C

Maybe maybe that's a good point.

B

um Yeah I think like Python and JavaScript go and Java and then.net I guess are like the five.

D

B

B

Those are the ones that seem like they'd, be the most important to have something by kubecon yeah.

D

Everything I.

E

D

E

All um they're all stable, no um and I I know that Dylan mentioned specifically he's not going to consider things GA until there's docs, just because last time not having enough docs for a ga release was a problem. I'm like cool.

B

Sounds good yeah yeah how that works.

C

Yeah, okay, so um I don't know: do we want just to settle with that structure and just get started there. um I mean I mean it's probably easier to to find the the Box. While going I mean we could put out a skeleton and work on that forever, but but maybe really saying like hey, let's take python. Let's get take JavaScript and Brook brick through that and then see how it's going right.

B

I I mean I'm pretty sure this, like this python example, for instance, was kind of built with the idea that it would be used for this, so I mean I. Think.

C

Yep yep no I mean rolling. The dice is exactly something where you want to have a metric or something like that and I don't know, do an average or counter or whatever so yeah. That works.

B

Yeah as long as we're covering the metric instruments and we're covering like the uh temporal aggregations and.

B

B

Do we have histogram section.

E

Yeah, that down below yeah another manual instrumentation side.

D

B

Yeah I mean I feel like that's.

B

I I think you've got it here, um probably and I honestly, like just start with python or whatever, and then write the doc. You know write the stuff out there and then we when we're duplicate, we want to repeat and then if people start repeating themselves, you know about Concepts and that gets moved into the concept section.

B

D

D

C

Okay, then maybe let's, let's take a few more days. If any anybody has some time. Okay, then maybe, let's, let's take a cornice in a month or something.

D

Something like.

C

That so, uh if you want to have something back down, maybe we should get started.

E

Yep yeah, that's my plan.

C

I think I have some network issues so.

E

Also uh in the uh in the interim here uh hi, uh is it Yana or Janna.

F

E

uh uh Jana I was told uh through my co-worker Mary, that Janna would be uh joining so I just want to make sure we had a chance to say hello.

F

um Hi, do you want me to give you any information about who I am and why I'm here.

D

Yeah absolutely.

F

My name is Jenna Avon um I'm, actually, a librarian uh I work in the academic sector, um but I am learning about writing docs and making a shift towards um moving to Tech. Writing for the tech industry um and Mary is a friend of mine from grad school and that's kind of a little bit of how I wound up here. It was I, know someone who is at least connected to the project, and so it is slightly less intimidated to jump in um and uh I.

F

Thank you in advance for all of your patience as I just kind of listen and do my best to help out where I get.

C

Awesome welcome. Thank you.

E

Yeah definitely welcome um so yeah I guess we'll probably go through the rest of the topics there, but then um I would say. uh I'm also happy to connect with you and sort of see, like you know, hey, what's like a like a reasonable thing to sort of hop in on I would imagine, certainly anything regarding like structure of information and like the the doc structure itself, like we always perpetually need to like improve that I guess.

E

So that's certainly a good place to jump in.

E

Cool, um do we want to continue through the the rest of the agenda here.

C

The the question is: should we do the end user stuff first, so I mean not not that we I mean we have around about 20 minutes and I. Think that that something we should do first right. Let me move this up real, quick, yes, yes, let me bring this up again, so I I hope I summarized this correctly, so um the the discussion started with um I.

C

Think you really you reached out wire slack or something and pointed to the blog post and they said, like hey, uh it's a little bit unfortunate that that the end user working group has to live out of a block. Wouldn't it be great if you have your own space uh within uh within the website, uh and then we then started this discussion and I. Think I made here a quick proposal uh to have something.

C

uh Let me see if this is still working yeah. Do you have something like here at the top alongside Community, something like end user and and there we could, let's say Precision the different things uh that discussion I think continued on the um did. It just continue on the on the pull request: I'm, not sure uh I, think we we settled then for this idea to say like hey.

C

Why not have something that looks a little bit more similar to the blog to select hey, there's a landing page and then or like the docs and then have some sort of structure here. On the left hand, side uh I think this also has a bunch of advantages, uh but then I'm wondering if we're not maybe even merge community and end users and have some kind of end user, slash community section where people can navigate. That's just a random idea, yeah just throw in additionally um but yeah.

C

That's what we wanted to discuss and it said I, it's not like. Maybe 20 minutes is required because that's a big change and um yeah how we should go about that. Can I make this picture a little bit bigger, oh yeah,.

B

um Just uh or enter on the discussion, I big, plus one, to merging this on to to putting this under community um like I, don't want or I, don't think it would be.

B

I, don't think it would make sense that I think having a separate end user sort of space in nav would be confusing, um or at least it would be.

B

It would draw an unnecessary distinction between like contributor, community and end user Community, because there's no reason like I, don't think that's anyone in the community should be empowered to contribute right, so I think we need to consider. We need to think of these. As you know, the same group with different needs um making that community.

B

Page structure have like the the blog sort of thing, would be: hmm there's Patrice, question I think, um but shouldn't be impossible um and then yeah like I, think. The other thing that would be challenging here would be like. Do you want those sub pages to kind of be more like top level? You know like content Pages, or are you okay with them just being like blog pages, because there is differences in what can be done with the with those templates, as is.

A

um If there's a space where we could like Rin and I or anyone else in the end user working group to update with um like current happenings um or news, that would be great. That's my well! That's at least my main um desire for the space. It's just like a Central easy way for us to update with news and like people can easily get like hey. This is what's happening this month or.

B

Yeah well, I would also really like to have a calendar on the in this full, because right now, I think we're kind of shoving a lot of stuff onto the hotel calendar, which is like fine, but it's also extremely busy on there now yeah.

A

And I don't think that, like I, subscribe to the hotel calendar and then was like oh no, my inbox, yep and I don't want to necessarily encourage folks who really want to just learn and not contribute to the project to subscribe to that calendar right now,.

B

A

B

A

It's overwhelming yeah.

B

We that's like a secondary thing, but I do think that there's there's this I feel like this Rises, the level of like custom implementation stuff that we are going to need to kind of engage with Patrice on, but from a like project level. I I would be really happy to see an expanded kind of community section that could include a space for the end user working group that could include a space for like hotel events, um both cncf ones and sort of Ambassador or community-led ones, um and also like contribute.

B

You know, contribution, guides and other kind of like hey. You want like, for instance, the hotel demo right. It needs a page where it needs a home on here, and the docs don't feel like a great place for it, but like a sub page in a community section does feel like a good place for it. Foreign.

C

That's exactly that I mean there's, there's more and more things that that somehow fit under that umbrella right I mean we also I, think had this discussion around I think it's now there, but I think it's not really how we wanted to have it uh uh the mission and and value section where we said like hey.

C

Actually, we want to have this copied over from uh consistently copied over from from the community people uh and that's maybe where is it and that's maybe another thing can can we like we do it with the spec, take some of the content here and and synchronize it live because I think not everybody is every time comfortable going into GitHub and digging all those stuff up. So so having those things.

B

Are always possible yeah.

C

Yeah exactly I know it's desirable. On the other hand, that's the difference. They're not no I, know I, know it's it's painful, I know, but.

B

um I would I mean I would probably I mean. How often is that changing though right, like I, think that some of this is just we need to make sure we have, you know, wait, you know like monthly or whatever it's like. Oh does. This is this need to be updated on the website without necessarily having to.

D

B

Things in because I think also there's some. You know longer term, like you can't maybe I don't know it's like formatting it nicer or like having. You know uh a video about the community.

A

Yeah I think something like that would be really cool, and maybe is a video is something we can pull together at this um Hotel. This um Hotel unplugged event.

B

But anyway, um yeah I think my two cents: let's can you all go yeah talk to Patrice, see what we can do about getting the community section expanded to have like a sidebar and subpages um so that people can kind of like flip back and forth between things, and then that can be where this stuff lives.

D

A

Is Patrice for context? Is that someone from the cncf or.

B

Yeah, that's that's they're a paid web dev contract they're, a paid contractor for they're assigned to hotel and a couple of other groups. Okay,.

D

B

uh They actually just got back from vacation, so.

B

And that actually, that if I don't know it'd be good to see if they could, if I could get scoped out and done by kubecon.

C

There's a lot we want to do by kubecon right, I.

B

Mean yeah I mean if you can't, then it can't, but it would be I mean I. Think the obviously. This is my personal opinion like metric stocks are priority zero here, like we.

A

B

Those done, um but this would probably be the second thing.

C

Now that totally makes sense so that, when someone clicks here on community that, especially maybe during kubecon, there's something what they can use to navigate and to find out like hey, there's some some.

B

Things going on these end user groups. Here's these upcoming events, here's.

D

B

You know here's because right now this is just like a bunch of links and.

C

B

That we can expand this out, like the actual content, development, isn't necessarily the hard part. um We mostly just need, like foreign, like an actual directory structure like an actual like content structure, where we have like yeah.

C

Pages and- and we could start with something like moving that stuff- that we have here- learn and connect and develop and contribute, maybe in in pages and then work from there and see like move, move those things gradually. So.

B

C

B

Like a landing page and then you have a secondary and then it's like your little sidebar and then you have end user, you know end user meetings or.

C

Whatever yeah and.

B

Then that takes you to a different thing.

C

I think that page here, right now is, is uh fed directly out of the doxy template. So it's always a little bit tricky to add it down. I think it's coming partially out of the config or something like that. That's a.

B

Special thing this is coming out of config. This.

C

Would probably be best.

B

Converted into but.

C

But yeah I mean you can throw it out and then just maybe do some copy and pasting and yeah.

B

C

D

C

Versus and then what what we would do additionally, is that we that we give code ownership. Maybe we have something like Community, approvers or maintainers, or something like that, uh so yeah.

B

C

B

If you had separate paths, if you had like a slash, Community slash, end user, then you could set up code owners for that um yeah. Whenever we get around to you, know more International, you know translations and stuff, then you could do it. For you know, whoever in China wants to spearhead Hotel.

C

Exactly yeah, okay cool, um but then it sounds like we. We have a kind of consensus on that uh so that we want to do that, uh and the first step is checking with Patrice how to have to accomplish that right.

D

So that's good to me, but awesome I'm.

C

Anything else on the community and use the space anything missing.

A

Seems good to me yeah.

B

A

D

Note that that's also where, like the demo.

B

C

C

Cool- let's go to the next one. I think that that's nevertheless, a quick one. uh We have two blog posts still pending. um Three actually um no worries. The thing I was just wondering about is we we have the announcement right now at the top of it, for the GC election.

C

um So should we keep it here uh for for another week, or should we just continue having having other blog posts and we've also promoted here at the top of the site? There was just a a question: I wanted to ask before yeah. We, we just push push new blog posts. On top of it.

B

B

C

Okay, that's good enough. Okay,.

B

So I thought about for a second and I realized: I, don't actually um I'm going through and it's it's 17 21. What are the other two.

C

ah Let me see I I added them a block label, so it's like um yeah, oh no, one of them does not have it yet.

B

So this is the one by me or do you want you to merge these.

C

I can I can merge them and do the date so I always like not having them like at the same day, so maybe even a schedule some some time between them.

C

So if you just approve them and go through them, uh then I said me or Philip can do the merging, let's say one tomorrow and the other one, maybe amid next week or something and there's another one uh here to the rust one, but but here I just reached out also to to the rust approvers to verify like is this because this is just a rust getting started um so I wanted to verify if this is um how we, how we want to have it in the block right, um but yeah I mean you can give this a look as well.

C

B

B

Yeah there's a decent amount of comments on that. um You might ping me when 1775 goes through some more reviews, everyone and I'll look over it then.

C

Yeah yeah, so the ones right now is is those.

B

Two are 1721 and 17.7 are good. Yeah.

D

B

um There's a future AI, it might be a.

B

I guess it doesn't matter because I was thinking, it's like it is kind of annoying that the oh did. We I think there was something with this like a way to modify those publicate, those dates when they actually publish through GitHub.

C

Yeah I think you can still edit them or sorry.

C

ah You mean that they they set to the current date when they're merged right, yeah.

B

It's probably something to do with templating, um because the problem is with Hugo if you, because you can like future date posts, but you have to rebuild the site because it'll skip things that have a date that is in the future.

B

I, don't know something to think about.

C

I mean right now we have let's say a blog post per week or something like that, so keeping the pr open and then going back to it once per week and then seeing like hey set the date of today and then just push it out. I think that's yeah.

B

That's fine, it.

C

B

um For now, yeah I mean we're doing a I will say we're doing a much better job about getting blog content on here and I. Think it's been very success. I think the migration's been very successful.

B

um This is another note because I was looking at this. The other day like uh We've as a Sig I, do want to give everyone their flowers, because we have really turned around the.

B

How authoritative open telemetry.io is this year like there were, you know into last year or like there were a lot of search, queries for open Telemetry, where open telemetry.io did not rank at all, and now it's like really hot. It's usually number one for most things, Hotel related, and it's also um like it's filling. You know all of Google's little like knowledge panel things are, are kind of picking it up, um which is good, so really good work. Everyone.

D

C

Cool, so we we have two more blog posts to come within the next few days and then maybe a third one.

C

um Anything else on the blog post, anybody's saying, like hey I, have another one. Let me let me put this out here right.

B

Now I do have another one. That's gonna be a bit, though it's about the new community maintainer role, I'm not finished. Writing it. I'd.

D

Expect also we'll have some blog posts around kubecon TBD. What those will actually be, but yeah I would anticipate at least one before one, possibly or maybe at most one before one during and one after.

B

More yeah, actually, on that note, did you want to should we write? Do like one announcement, blog post.

D

B

Because I think it.

D

Makes I think the main announcement would be the the demo project right.

B

Well, the the metrics GA and demo right, yeah, so I, I, think and there's probably some others. There's other things here and there, but I think it makes sense to kind of have like one thing and that's it's easier to get one thing posted on the cncf blog.

D

B

Easier to have one thing that we kind.

D

B

D

Should probably have one before the conference just saying here are the talks like here are the places you can.

B

Go right: we need to have the so yeah so pre-conference. That's what I want to note this um or just note this so that I can do it later. uh We need to have an announcement. We need to have a blog post that talks about like hey kubecon preview, um which is like Hotel unplugged observability open observability day, um and then the talks that are going to the maintainer track or the maintainer track stuff, um and then we need to have the one for the 25th I. Think that'll be the actual like project.

D

Updates project.

B

Updates the one that we, the Press, wouldn't be the press, one basically.

D

And then probably one after maybe.

E

B

I think yeah recap: if we don't I, think we'll have like there'll be enough people at these various things. So once if someone wants to take Point, um writing it or I, guess technically I think technically it would be my.

D

B

Take a point on radio.

D

B

um Just figure this out, but yeah we'll have a wrap up blog post um after kubecon.

A

Yeah and um for reasons of the fact that we have not been very active as a project on the cncf blog I'd like to think about. Maybe.

A

As well so we start getting in a habit of getting content. There.

B

Yes, well now we have um me to help sort of proxy those things and work on post post.

A

Yeah and I'm happy to help as well. I've got access to service desk and all of that.

B

Oh perfect, then great: we can both work on it. Then. um Actually, let's um we're on slack together, yeah.

A

But there's a lot of tiny projects that are like every one of their release. Notes goes to the CM, CF blog and, while I don't think, we need to get it that.

B

Way it's a bit excessive I do think we need like monthly, yeah, I, think monthly monthly Cadence would be good. Yeah.

A

That's exactly what I think is something monthly there that could alternate between bigger ideas related to observability and open Telemetry things like kubecon Recaps thing like this isn't fundamentally new feature that we've released and it's really interesting.

B

Yeah I think the biggest one yeah, the big one or maybe even as a regular part of it, is just like up to you, know new releases and important things from the sigs.

B

um Because that's one thing that isn't necessarily covered super well right now, like I, know, there's been some interest in.

B

Oh um someone was I forget who was asking about this, but like tweeting, whenever there's like a new release of The Collector or whatever right, we can, we can figure it out.

D

So one cncf blog post for every point of update. That's what I'm hearing exactly joking.

D

C

Awesome we have two minutes left one one topic um semi-related to the first one, but we can also take this offline, so we reached out to C plus plus Sig around uh the dogs, and um they want to move as well over to open Telemetry. I o um long discussion. The the only thing that that stood uh is that lalit said something like that during the maintenance meeting, uh the the docs team would come up with guideline processes for publishing documents, um I'm, not sure Philip.

C

Maybe you remember what she's referring to it's just a metrics things we were just talking about, or it's.

E

It's related to that there was another like discussion. Sort of um uh this is this is mostly myself and uh damn it uh Tyler from the from the gosic um Tyler kind of had the wondering thought of like oh well, you know you know there. There can be issues of like docstrift and, like you know, some stuff, potentially not compiling anymore, because, like some apis change or something like that and like how do we deal with that?

E

And uh you know, would it would it be easier to like you know, standardize on on the whole sub modules thing and I'm like well?

E

um uh There is a reason why we pulled a lot of the languages into the hotel repo, because the whole sub modules approach didn't actually work out and people were not updating their docs and it was actually better to just like have someone who spends time on the docs themselves, just keeping stuff up to date.

E

um I, don't know if we like arrived at a broad consensus that everybody shared and so I think, like the takeaway that I had there is it's. It's still like. If, like say, the C plus team wants to be very active with writing their own documentation, um that's great! We should really encourage that uh and if they're like you know, hey, we want to have like this stuff be on the website.

E

But then this other stuff like be pulled in Via sub module like I, would try to encourage them to standardize on just one of the other for all things docs, but um with a preference for living in the docs repo. But um you know we should we should Empower them as much as possible. I think um and whatever process we have established Now isn't always going to be process. That's there forever, like I mean like the the the thing that I brought up is like.

E

Also tracing is stable now and so the likelihood that there's going to be like wide sweeping changes for a given language if they've actually declared their API and SDK stable, is very, very low and so like the need to live close to the code. That makes those changes is far less because of that. um So that's like another another Factor there. So I could see a world where someone's like. Well, we want to actively develop our docs in our own repo and then maybe do like a doc dump or or something um I feel like.

E

We could accommodate pretty much anything so long as it's. It's not a um a Rube Goldberg machine that we have to set up to get it all working.

D

C

Okay, so it was not around like hey, we create a set of guidelines and then we have to go through all the maintainers and get their approval. It was more around like a. We continue to do, allow a lot of things you're doing, but but here's some best practices. We recommend.

E

Yeah and so I mean I, think I think we should. We should still chat with them again and um yeah. You know, like our.

E

Our preference clearly would be that, like at a minimum, the tracing docks would live in our repo because it's stable across the board and then API docs are you know, published and wherever they're published to and then linked to um and then four metrics we can go to them with sort of the same structure that that we would with the others and say, like you know, hey, do you want to approach it this way or do you want to just table that for now and and like get what.

B

E

Got already into a stable State, um yes,.

C

I'm cool feather.

D

C

Okay, okay, no no I! Think then then I understand I, think I mean eventually that it was was what we discussed in the beginning. We want to have similar talks for for all languages, but yeah at the end. The six themselves have to decide how they how they want to go about it right: um okay, cool yeah, I, try to to come up with something the thing I want to do. Midterm is really have to say: please add some C plus plus getting started manual.

C

Instrumentation exporters Etc, like we have it for for the other languages right, so um yeah, anyways, I, think I also need to jump to the next one uh or anything else, not from my side.

D

No I think I got everything here. Awesome.

E

Yeah uh I also just wanted to say uh Jonah. Thank you for for joining. This is definitely jumping into the deep end immediately here.

E

um uh If I don't know how how familiar you are with um uh uh GitHub and sort of the DOCSIS code approach to to writing docs, um but we are starting to standardize on issues there to like label things as like: hey these are good to start with um and then there's sections of our documentation uh that might be easier to hop into compared to others.

E

um So like, for example, if you're like hey I, want to help with documenting metrics for golang. Well, that's probably going to be a lot harder because then you know there's like a lot of moving pieces.

E

Some stuff may not be like quite up to date yet, and there and like you, got to kind of be really tuned into like how it's going with go, whereas if you're, like hey, there's like these concepts for open Telemetry that are language, agnostic and stable like and they exist in a specification but they're not like really human, readable figuring out a way like to like make that stuff human, readable and part of our conceptual docs is like far better to start with and stuff and so um I think.

E

If you're interested in in contributing, we can definitely like tee up some things that are, that are much more straightforward to to bite off. I. Think yeah.

F

I I appreciate that I. Definitely a lot of obviously we've been talking about so far in this meeting is a little over my head. So it's been good to just kind of like listen and soak it in um I.

F

Have a lot of transferable skills for some librarianship and information organization and Records management and Technical instruction and all of that stuff, so I I think there are areas where I can contribute I think to say um what you're suggesting around um human readable uh technical language is, is probably a great start for me, um but yeah I'm I mean I I just want to learn, that's really that's where I'm at, and this seemed as good a place to start as any other one. At least I knew somebody who was remotely connected to it. Yeah.

E

Yeah definitely um so, let's, let's definitely stay in touch and- and you know, there's there's no shortage of things. I think that we could have on the hotel side that um could stand to be improved, even just from like an information architecture, standpoint and so um figuring out a way to.

F

um We're still recordings I, don't like super, want to talk too much about this stuff. uh Do how should I follow up with you.

E

um Let's see, are you in the cncf uh Slack.

F

E

Yet, okay, um that would that that'll be important, because then there is a channel there, uh Hotel Dash coms, that's that's the the channel for this group, uh Comes This, Communications website, um docs, blog um and yeah from there that that's how we could we could chat and sort of figure out good things to to do there and then I I think that this could probably also help Severn and I like better identify like things that are like good starter issues versus not good starter issues. Great.

C

I I put there's a short link, so there's slack.cncfio, you click there and then you I think you can register with it and uh the channel I think is called Hotel, cops, I, think or something like that.

C

That's the two things you need! If, uh if, if you run into any issues, uh yeah reach out to us by any other means, um we can can see perfect awesome. Thank you very much. Thanks.

F

C

Thanks for being here and talk to you, then in two weeks sounds good awesome, yeah, bye-bye.