Open Telemetry Uncategorized, 7 Apr 2023

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2023-04-07 meeting

Description

OTel Rust SIG

A

A

B

I think I have some audio.

C

B

Hello, can you hear yep, okay, I just joined from my mobile phone, because my laptop doesn't like Zoom for some reason.

D

You wouldn't be the first person.

B

Yeah I usually joined from my laptop, but many times. I got some complaints that my audio was very much uh very feebled. So I have my backup plan, which is to use my phone nice.

D

uh I think we're.

A

D

A

D

um Other than the topics and the minutes uh did y'all have anything to perfect.

B

Yeah I would mostly be here just to say, hello and just meet people. Folks, yeah, that's my main goal: I'm, hoping that we'll be able to continue uh having a recurring meeting, yeah weekly or once in two weeks would work.

D

Okay, yeah, it looks like in the well since I think everybody that RSVP is here.

D

um Maybe we can go around, uh introduce everybody, uh I'm, Harold, Dost, uh relatively new to well semi-new to the to the to the rest part of the project, but I worked a little bit on the specs did some other stuff, um but uh had been somewhat dormant uh and tried to get more recently uh acquainted with the with the project.

D

um I, don't know if you want to go next.

B

Yeah I don't know who else he said, but I'll just go ahead: yeah yeah, my name is sijo I work for Microsoft I've, been active from uh like the beginning of open Elementary like late 2019, early 2020, uh but mostly restricted to the open telemetry.net six. uh So that's my background in open, Elementary, um like very new to rust, not just open, Elementary Rush like rust, essay language is completely new to me.

B

So I work in Microsoft and there are few teams within Microsoft who are like uh heavily using rust and they uh asked my team support for shipping like open Telemetry rest, so they can start using that in their production. So you can expect to find me here for like very active in the next. uh Like maybe I, don't know six months or whatever it. However, it takes to make sure like we can uh get things out uh and you'll also hear from many of my colleagues teammates as well uh yeah. That's it yes!.

A

Hey hi guys I can uh I can go next. Sorry, my I think camera is a bit you're.

D

A

I'll just switch it off yeah, so hi I'm, lalit uh I have been working on open, Telemetry I think it's not two years uh working as a right now, I'm working mostly on open, Telemetry, C, plus as a material code.

A

Yeah I think it's working for Microsoft and I. Think as CEO has said, Microsoft has some interest, basically in open Telemetry rest. So probably we will be supporting some time on trying to contribute uh some some of the components per rest. So.

C

Totally new to rest.

A

I think most of my experience has been on C plus, so please bear with us I mean in case. We contribute something very name and something which we see so yeah so I think probably we'll learn uh along with yoga A7, as we continue motivating to rest project.

C

Okay, uh I guess it's my turn.

C

I have been with this project, I guess uh ever since Julia introduced the first Matrix implantation I'll still review it. So it's been two years I guess: um I haven't been contributing as much lately, but I plan to run back up. um Yeah I've been mostly working on wrestling. This project I, don't really use it on my company settings, so I can't really say I'm a total expert but I get I, guess I, guess some experience working with rust.

D

um So I guess with that, uh do we want to try to set up a regular time, there's currently two slots that are currently already set up for us, I guess when, when you and the others were uh we're I, guess starting it out, I guess the last one happened back in.

C

A

D

A

D

um So is is 2 p.m, UTC and 11 p.m. Tc good for people I mean like at least that that gives uh you know for people that want to join different time zones. It's you know more favorable back and forth, because it's basically almost the opposite.

B

So, where are you based on Heron, like? Are you in the uh European Time Zone yeah.

D

Yeah I'm in uh I'm in Paris time, so, okay.

B

And Shong Yang: are you in the uh us like this.

C

Yeah I mean Seattle, so okay, East Coast.

C

And this thing I think what happened pretty decent contribution from European side. So maybe there's a decent interest from European side. So.

B

So uh like Harold, can you say what was the two proposed times again like UTC, 4 pm or UTC 2 p.m? It.

D

Was it was UTC 2PM and UTC 11 p.m, and there were um they were basically alternating every other week. uh If you look in the notes, it's the Google Calendar, that's out there. This is like the official one that basically every Sig that has a meeting uh is in and the rust one does exist already in there and we can choose to just follow what's there.

D

If we want to continue with that, if we want to do something else, then we can just ask them to change it uh so that it's at least so it's at least reflected properly.

B

I'm, okay, with their time like I'm, very flexible with my timing, so I would be like the least uh worry for it, because I wake up very early, so timing wouldn't be an issue for me: yeah I'm, okay, with sticking with whatever is there in the calendar. So that's every alternate week, 4 pm in my time and 7 am in my time so yeah. That's totally fine for me. Yeah.

D

I, probably wouldn't be making the 11 pm UTC ones, just because that's uh 2, A.M or 1am 1am 1am.

C

D

Depends on the time the time here.

B

If it's uh like, if it's too late, like maybe you could, we could all like agree to not have that one like just stick with the time where most of us can make.

C

E

Right yo 7am is kind of stress. I'm not gonna lie.

D

Well, and that's that's: why I think it's it's good to have uh have the two alternating ones, because it's like at least gives uh at least gives people to the ability to to do a little bit of both I mean not everybody's gonna, be at every meeting anyway.

B

D

At least, if we're able to make incremental progress, I think that's better than nothing and like.

A

D

I have no plans of being at a 1am meeting, but.

E

D

E

Is there like a large Cadence for uh like the number of people in.

D

E

D

Mean this is this is the first meeting outside of since 2021, so we don't really have a Cadence at this point. I.

E

Think at least like what you remembered like two years ago, I guess like or.

C

Was there just not a.

E

Lot of yeah, okay.

C

Yeah, so basically, whenever we try to start this meeting so for the first first few instances, people are drawing, but it gradually people just stop drawing I. Think that's the one in 2021. It's just me Julianne and one other folks like us and there's not really nothing, not really something to talk about, because with that Cadence it's not. We can't already do the three suit triage. So just talk about General, High, Level topics and start to attacking issues all around.

E

Yeah the reason why I'm asking is because like if we don't know that there are a lot of contributors right or like if that will be okay with that time and like you know, we do a 4, P.M or sorry uh was it like 11, um like 1am, for you Harold um and then like no one shows up to that meeting like. Why do we even have that kind of time slot in the first place right like we? Don't we don't really know yet right so like?

E

Maybe we should just try it first and then um we do need someone to go there to see how many people show up, though right in the first place right sounds.

D

Like there's only like one person here for that.

E

I can I can make that I can make that meeting. You know um as.

C

Well, I guess would it be helpful if we start some kind of survey on the slack Channel, maybe because no that's our unfortunately Communication channel to gathers um like what's people's uh what a time that people's comfortable is to join the meeting also I, do think. 7Am is a little bit too early. Personally I, you really don't wake up until now,.

E

Like this, this one right here is like the perfect time, but you know I know enough.

C

For everyone I wonder if we should push push 7 A.M a little bit late and push the 4 pm one a little bit early. So it's like uh it's come.

D

Towards some kind.

C

Of news, yeah yeah, because I do think that, like the current time, is a little bit stretch either it's too early for U.S folks or too late for your Europe folks. So.

D

Yeah yeah cause like right now: I mean on a Tuesday I'm, okay with 6PM, but typically I. Don't prefer it on a Friday I already got dressed up, ready to go out so like after.

A

E

D

Over I'm out of here.

E

Is 6 p.m? Right now, for you like yeah, okay, true.

B

If you're, okay, two stays on this day, maybe we could pick Tuesday.

C

D

Yeah I mean that works that works, maybe so we could uh so we're saying is: maybe keep it at this time that we had and just make that the new time for Tuesdays um yep, okay, so.

B

Every Tuesday same time or like we are flipping or rotating between two schedules,.

D

uh uh I think yeah I mean I'm fine with doing one of the time slots as uh what is it 4pmtc? Oh okay, so.

A

It would be 4 P.M, 2.

D

Pmbtc well or maybe.

E

4 pm UTC is right now, like 9 00 a.m.

E

uh Utc yeah.

D

That's right: yes, okay, yeah, yeah,.

B

Maybe we can stick with uh 4 pm UTC every Tuesday at least to begin with, and then yeah.

B

Was it you who made the zoom link or.

D

um No so there's there's like a couple of meeting rooms and they alternate so there's like zero through four or something like that. Yeah. um So I think we just need to make a ticket with the community channel to make sure that that that's clear.

B

A

D

They'll take care of the making sure that's aligned.

B

Is our like planned meeting time for the next at least two three weeks.

E

Yeah, did you guys do like intros already sorry I.

D

Was late? Yes, we do yeah. Sorry, we couldn't. uh Oh.

E

D

That no no you're good! uh If you want to do a quick one, oh.

E

Yeah hi guys I'm I'm Layton um I work with CG on the lit um uh from Microsoft uh yeah. We just taken recent interest in in Rust, so um never done it before you know so, I'm still like learning and going through it, but like I'm I'm, happy to see that um you guys are coming running across like the same problems. We did um oh I'm a maintainer for OT python, um so I'm excited to kind of redo open. Telemetry again you know it's like I.

E

Think we've all been through that process, uh but yeah like like rust is uh is new for me. So uh um hopefully you guys for forgive my. You know dumb questions in the future, um because it is quite A, New Concept for me. So.

D

um Continually.

E

Yeah man, rough life, we chose the weird profession man, so um at least we get to sit at home and just talk to the to the camera right. So exactly yeah.

B

So, just for others, knowledge like light and let it myself we are all part of the same, uh not necessarily the same manager but at least the same organization, and we all have been working in open, Elementary in different iverson.net, let it in C plus plus light and in Python. We may have, like others, joining like on a part-time basis, but at least three of us would be expecting that we'll be like working more than 50 of our time in this effort. Okay,.

D

Nice, um so it's something that's not necessarily the highest priority topic, but I wanted to bring it up. Was um the cleanup the of the user groups?

D

um I, don't know what the typical, uh what the typical procedure is on this um on other on other contributor groups, but like uh just in the typical security list, the least uh was it uh security at least access.

A

D

These privileges, that's what I was looking for, is uh we still have people who haven't contributed, haven't uh really been active on the project um in a couple years now and so I don't know you're the I think you're the only one that actually has well that's here. At least that has access to make that change.

C

um Yeah I mean I can try, but I think my access compared with Julian has still is a little bit different. So we may have to talk to uh like the community Channel ask for help so, but I'll just give it a try.

B

C

B

E

C

Yeah I'm one of the swimming units left now.

E

Got it sorry, they.

B

Make you feel removing some? Please listen as the Embry test status, because there was a clarification last year that, if you're removing someone list them as Emirates maintenance or whatever.

C

B

C

Okay so we create a new group or we just list them as the m s.

E

You just list it as a Emeritus approver.

C

E

Maintainer in, like your contributing document, yeah.

C

Oh yeah yeah our data.

B

Okay, if that's the case, then all you need is just remove them from all the privileged user access groups.

C

Yeah I think I won't have so I will give it a try if it's not working I would just reach out to your community I. Ask him to do it for us, okay, cool.

E

A

Was hyped to see that.

E

There was like eight members in the rest, approvers, you know like people.

C

Come and go like yeah.

E

C

Come and go there were a lot of people contributing to this project like some of them just uh I, don't know exactly, then they just stopped doing that, but it's okay. We still appreciate their contributions. Yeah.

E

Yeah, no, that's awesome. Yeah yeah, like every time like we start a sick. It's like three people and it's like all the maintainers. You know it's like all right, guys, I guess we're doing bugging again. That's like.

C

That's gonna be a song yeah.

E

So uh also is that how you pronounce your name, properly, yeah yeah, awesome.

E

uh Any other topics Harold.

D

uh Well, actually, you could you had brought up the contribute, repo I figured that was going to be a smaller topic than the path to GA, so I put it up higher. Oh.

C

I saw it I thought I put it in one place,.

C

Okay, let's discuss contributorial first I guess um so, as we have discussed in the issue. I have we have issue regarding contributory, but we should um which have a contributable to compare any I guess non-official packages.

C

um For example, we have a data loggers, but I guess one of the topic I want to mention is: what's the first, it was Guidance with someone contributor, that's kind of unofficial package to us. For example, the the dinosaur perspective is contributed totally by one individuals in one PR and we accept it.

C

um But if we find it's hard to maintain going forward because it's extra package when you update something update along in all those official packages, so is it recommend to keep them into comfortable, comfortable, Ripple and assume at least some level of maintainers Maintenance or we're just asking to have their own GitHub repo holds the third party packages I.

D

Mean it's yeah. Repo concept was that basically like this is something that can exist and that I mean and- and you guys can correct me if I'm wrong but I I for my interpretation is like you have the official stuff, and this is like, like top level of support and then everything in contribute is kind of like you can use this and depending on the particular plugin uh you know it, it might be at a good level.

D

It might be stable that sort of thing, but you have to look at the specific uh plug-in because yeah, because it varies.

C

Yeah this makes sense. If we do have contributed version, we should make sure we communicated that that, basically we are using country, then it's they. It won't be as good as the official one in terms of Maintenance everything yeah.

E

I do want to point out um so for at least for python we have a core repo and a contributable uh there's a lot of headache that came with um synchronizing the two repositories as well as like, uh like testing.

E

Whenever there's like breaking changes and stuff like that, um we have to do like very annoying build systems in which, like um we check out the secondary repo every time we pull a PR out to test, because if I make a change in the SDK or the API right, it might break something in the instrumentations right yeah. And then we test it every single time. There's a change, so we check we physically check out in the workflow. The contrib repo run all of the instrumentation tests against the new change, and we also do the reverse.

E

So if we put a PR out for the contribute repo, we check out the the core repo and run the API SDK tests on the instrumentation changes. So there are a lot of pros and cons for doing this.

E

um I'm going to tell you that the maintenance burden is is horrendous like like it's really bad yeah, okay, um uh so um yeah! That's just my two cents, the the the pros is exactly what you said, though. It's like. We have much more uh um like strict guidelines for like what we maintain uh like. If some random person came with an instrumentation right, um we could be like okay, we're not accepting any more instrumentations or uh we have the kind of deprecation life cycle that um a lot of other languages are doing.

E

It's like, okay, if there's no uh PR's on this specific uh instrumentation, like we deprecated after like six months or something like that, and people could still use it if they want. But you know keep in mind that it's going to be in some sort of beta form right and it's much more powerful like easier to do that. If you have a contrib repo um but yeah, those are pretty like the cons are pretty bad. I'm, just gonna point that out yeah.

C

In terms of in terms of a cross field between those two repos I think for control, we don't have to use the mask Branch. They use the released Branch, it's just gonna I guess it doesn't really eliminate that it maintenance experience, just push it down when you try to release a new version, but I guess that's. At least today, when you have a PR, you don't have to worry about the brick stuff. Bricks are the Ripple.

B

But this is not a long-term issue, because it's only an issue. If you keep breaking things once you do like one daughter, then there shouldn't be any breaking changes, so that should not be an issue at all like it would be an issue in the very early stages, but like at least in.net, we could barely once we did 1.0. We never had to worry about any anything in the country, people being broken, because after one daughter, it's all like incremental changes. Only.

E

Well, it's not that short term, because a lot of instrumentations won't be stable right. So.

B

No, you could go the other way right. The EPA and SDK would be like one city, stable, instrumentations or exporters, or anything which depends on the core. Api and District label. Never need any need to react to any breaking change.

E

Is not sometimes I think sometimes.

B

Oh, then, that's a something which we need to fix, because if, after after 1., what after stable has occurred, if there is any breaking change, that's a big concern, so I I would say it should not occur. If it occurs, then something is wrong with the process, because it's not just the instrumentations we write there could be other customers who might be broken.

B

So what dotnet did was uh we moved all the instrumentations exporters which are not officially part of the spec into the country, but we decided to keep like two or three most popular one in the main repo itself, so for uh in case of dot net the asp.net and HTTP, they are the most popular instrumentations. We kept it in the main repo just to address. The issue like and suggested like to keep a good set of integration and test coverage, because, if they're in the same report, it's very easy to do that integration test.

B

uh So if you are not breaking those two core, then chances are pretty low, that you're breaking any other instrumentations in the country. But yes, it's theoretically possible until you reach the stable state.

D

Okay, so it might be something that we defer to like. Okay, once we have let's say tracing and metrics stable release, then consider doing it, but until then might might be easier to keep it all. In the same.

B

B

Matrix API right so if there were instrumentations in other reports, that would have been broken uh yeah, but there is no instrumentations yet interest rate. I I did not see any instrumentation libraries similar to other languages in open, Elementary, rust.

C

To be single repo themselves, if I remember correctly,.

B

Oh, like you, have a dedicated report for a single instrumentation package itself,.

C

No I think like for instrumentation for popular HTTP framework or JPC framework. They already have their own GitHub repo and they just maintain themselves. Nice.

E

Okay, that's a good idea too like so then we have the maintenance burden on the contributors. So.

B

Yeah, so this was only or like discussed like long back even in.net like we decided we'll have a country repo, we'll start with that, and if, like some components grow like really well, they could spin off into their own repo, uh but it never happens. It's still in the same, like uber report, which has like 10 or maybe like several dozens of contributions right now,.

C

But if you have uh yeah.

B

No, if they grow big, like we could always spin off to Dedicated reports or ask them to take it to their own, like if it's owned by like some companies, they could take it to their own.

C

Yeah I just have a list of uh repo that either maintained by ah or something by other people and they sometimes people just add their Ripple there. Oh.

B

Cool, that's pretty cool yeah, so my only uh ask would be like if you have a, if you do not plan to do a con group reported list like I mean before we do like one dot of uh stable of tracing or metrics I want to know like if it would be okay for uh the current maintenance to accept contributions from us, which uh it would be like not a specification related PRS which you'll be sending, uh because we have a requirement to export to operating system.

B

uh Primitives like etw in Windows and user events in Linux. It's mostly like open, like 90 but like 10, could be like somewhat uh microsoft-ish, not really internal. uh So we want to make sure like we have a place to contribute that and keep it ideally in a country report or if, if you plan to keep the contributions, okay in the main report, we are okay with that as well. uh It's like mostly open like like each level is open.

B

Userability is open, it's all like open, but it's not something which is written by the spec. The open, Elementary specification does not have an exporter to these OS native things. So that's, why usually is hosted in a country paper like at least for C, plus plus and four.net? Be a hostility name separate triple, uh but we are okay either way like as long as we have a place to put our code. uh That would work for us either in the main repo or in the control.

C

Yes, I think like we'll talk about country proposed. Only what we were discussing is how we organize those comfortable components like the concept that you already have like contributory pulling or main reports, just how we organized call whether in one report integral so it's fine you're, accepting any code that should have landed in comfortable with pricing it. Just we want.

D

B

C

Them maybe from the main report to another one, but accepting it shouldn't be a problem. We have already have a open, Elementary country great, but depends on the complexity of the exporter. Maybe it was to it like publish a separate query.

B

Okay, yeah, that would work for us, I mean at least for now like in the future. If you think it's going to be easier to have a separate repo yeah, we can do that as well.

D

To summarize is just gonna say like we can uh yeah.

C

We can accept it basically now and we have discussed where we put those comfortable components, whether in main report we're going to discuss later, but we for now.

B

Yeah, it will be at least several weeks before we actually be contributing actual code, so we'll have plenty of time to decide that yeah.

C

The issues issue we have on contribute contract, Ripple.

B

D

Okay, uh so I guess the final topic, unless anybody else has something that they wanted to bring up would be trying to get tracing to. Ga um I know that uh from the last the last meeting that was had there, the last League meeting that was had there was you know. Talk about, like um you know, as you mentioned, need some more test coverage.

D

um We want to potentially run through the rust API checklist uh yeah. Some some other things like that uh dog food survey. I did not see no open, Telemetry.

C

Oh, the dog Force server, basically is a questionnaire for customers, say there's anything you want has to improve before we say what everything else and uh across the apan library checklist is the DraStic thing. It's a published by either the official grass language or some very powerful, very influential organization within us, basically give your guidance on what what's the address. Rusty API should look like, and my primary concern I think I add those two back in a day.

C

My primary concern with current state of tracing is: we do have a lot of delay test coverage, but I don't think we have any if or much integration test coverage.

C

I do I wrote integration test on Jaeger, but I think we should probably add one for open, Elementary collector, so I thought it should be sending about a collector, like maybe a container and then start fire up test instance of open inventory, rust and then certain Choice expense to The Collector and then just carry the character somehow to see whether it looks looks like what we expected to looks like. Okay,.

D

So that might be a good, uh a good additional ticket that we can create.

C

Yeah and then in terms of API, comprehensions I think we should probably do it review it, uh one by ourselves and then meeting together to save their assigning missing pieces.

E

Yeah some something that we found really useful when we're when we did our tracing uh before we did our chasing J was that we went through like the spec compliance Matrix um uh that exists in the spec repo I. Make sure that, like you know, the language just implements all of the the boxes, um and this is this- is a manual process, so um something we did for metrics GA was like.

E

We actually even got like a GC member or a TC member um to actually go through our API surface and like they found a couple of issues and stuff, and then they gave the go-ahead.

E

Okay, don't know if we want to have to be that intricate, um but given that this is your first ga you know might be something to think about not GA.

A

Sorry uh stable.

E

A

Yeah, probably just go ahead: I think we did the similar thing. When we had the tracing GA, we did go through the compliance metrics and also we went for the review by the PC member and then subsequently for the Matrix I. Think we had enough confidence going to be a compliance, Matrix and I. Think we put you know: Among Us, you know all the maintenance.

A

We could take a decision when it's the right time for the engine, so yeah at least for the for the traces I think we did go for the review with the TC member and it was very helpful for us, at least for the C plus community.

B

So yeah, if you can get a TC member to review, that would be always helpful because they usually have bigger knowledge that about the spec than all of us, yeah tracing. We got TC members to review it because we were the first language to go. Ga.

B

But for Matrix, no, we didn't do it like we just decided among all the maintenance. Okay, we are good to go so like for us. This is it's the first time you are doing a 1.0 stable. It might make sense to get someone from TC, but that that's okay, voluntary thing there is no process or anything we'll just ask someone to review. uh Whoever has some background interest might find it easier, uh but yeah. We could always uh request someone from DC to.

D

I do currently have a stability tracing ticket to try to solve it, or at least try to track it um and there's a milestone for any like tickets, that we end up creating um one of the things that maybe useful is if we can because I know, one of the things that we did with the spec was like. There was a whole bunch of issues and it was like okay.

D

How many of these do we have to care about before we go stable because, like right now, I think there's 20, including the stability stability tracing one uh just because we weren't stable at the time and.

E

They marked as a trace the area Trace tag yeah. Those are the issues you're referring to.

D

uh Yeah, in addition to that, I did actually occurred in Milestone just for like stability, because there might be certain things that might be Trace related, but I don't know, maybe they're related, maybe they're, not um I. Think I started with basically everything that was in Trace. There's 26.

D

things um yeah because, like consider, deprecating, open, Telemetry, Jaeger, honestly doesn't matter for if we stabilize the APA.

A

D

Because right, it's an exporter, so we don't actually have to care about that. uh But like there's this one of how to fix breaking changes with with Trace install batch foreign.

C

I think someone replied, but the opener didn't really respond.

D

Yeah yeah, it looks like that, but yeah.

C

B

D

That's an example is like, maybe maybe it was, maybe it.

A

D

This case I, don't I, think you're right I. Don't think that necessarily was is a breaking thing, but like the propagators API, that's something that we should probably have um right for stable, because just just so that we're in in a good spot there, but so.

B

What is the plan to release the APA and the sdks table? uh What about the exporters? Will they also be marked as stable, along with the APA and SDK the at least the otlp exporter?.

C

Yeah I think for for 1.0 or ja. We should at least have one or two exporter that defined in the spec as official support exporter to be stable to otherwise doesn't really make sense if you have only instrument it without exporting it. A single trp is a good candidate for that wait, and otherwise Jager actually a year ago, is the most used one and then no trp one well yeah going back to my parents point. If we want to release anything at the world comfortable, you have end-to-end integration tests for it.

D

E

Zipkin very popular in this.

C

E

C

Have um so last time for official tax distributor so to speak, so if I actually say how many times those reports being downloaded, unless uh let me send it to so basically there's open Elementary, you can just basically say open, Elementary Dash the king I mean the download number may not be accurate, I. Imagine a nice portion of it is automatic download by some company.

E

Yeah yeah yeah, but.

C

I mean it gives you a indicates.

D

Yeah there does seem to be, and it's like, I get it from a communication standpoint, but one thing that's been a little frustrating is that, like there was somebody who wrote a blog post on like how you can use the rust toolkit with uh the Jaeger exporter and the person's like, ah but that's not the official one, it's deprecated, and it's like yes, but this is what people use like on the ground when it comes when it comes to Brass text. This is what people are using, don't like.

D

Don't try to like force things down people's throats when they're like actually trying to solve problems right now,.

E

There's always gonna be those people, man, that's true.

D

D

But um yeah so I uh to round up I think you said: API SDK will say: yeah, otlp and Jaeger. If.

C

D

Are stable, then that would be good and we can worry about stabilizing Zipkin afterwards, because its priority is a bit lower. That sounds.

E

E

I have a follow-up question about the the stable. um Is there a rough plan on how um we will release the other signals.

C

Already yeah, so the current place when we say one point there or explicit: it's only applied to tracing and even after 1.0 people should still expect breaking change on logging and Matrix and when Matrix stable enough, which is really it's two pointed oh okay, nice and then so. Ideally, we should will be a swing point there for we stabilized, obviously I.

E

Mean yeah: that's a good plan, yeah yeah. What we did for python was like. Oh all, the metrics are just additive changes, so we're still like 1.6 or something like that. Even though uh metric metric signal was like unstable for like a year, so yeah 43.0 is like weight cleaner uh more than two years man. That was ridiculous yeah. So that sounds good yeah.

B

So rust went through the same process as other languages for metrics, because Matrix was already and like 509.net following the old spec. But then the spec was like completely Rewritten and we all like throw away.

E

We were like one of the.

D

First ones, that was sad.

E

D

E

B

D

I mean you guys were the you: were the yeah you're, the the templating language.

E

Right, the pro yeah they even asked me like they were like make this, and then we can have something to sell the community like okay and then they use that as like a thick guidance and I'm like nah. This sucks.

E

Yeah I mean these jokes.

E

Okay- um oh sorry, just also one more question, oh, is, is the best way for communication through slack like if I have any questions or anything on how to do stuff.

C

Yeah uh we'll have mixed attack our official Communication channel myself for last week or not last week,.

C

Because, oh stack is not already open enough um in.

E

C

Of the members mine so I see but yeah we'll make this step and move to start.

D

Yeah we have a foot in both and I know that there's a couple other cigs that also still have their getter. So the one thing is right now, I noticed on the Community page. It only mentions the getter, so I was going to add the slack and say slack and get her there. That way. At least people know like hey. We can go to both places.

C

Oh, that's, that's not the end of it. Like you, ask us to the program. Language especially, is night to use Discord as Communication channel, so you say that that's the official language so over there is it the most popular job you say uh asynchronous framework. The Tokyo is there so sometimes well also just looks through their Communications. If they're anything really do open, Elementary and there usually is like in the tracing channel of the token server, ah damn yeah. So that's that's that.

D

Okay, so I have to join the Tokyo. Discord is what I'm here yeah. We can go to Tokyo.

B

You have all three coverages, like slack Jitter Discord,.

E

All three of them open and need like different monitors. Oh God,.

C

It's funny because my the programming is server is right. Next to my gaming server, it's like so I was just stepping through one server yeah.

E

You start asking rust questions and like you're, gay.

D

E

D

E

What is this sorry yeah? It was all yeah wrong, rust. Yes, nice, nice.

E

um Is there a way to kind of like um I I know we don't really like uh use the tags for this? um Is there like a way to see if there's like easy issues to pick up.

C

There are some good first issues there like we have this neighbor and you have to help want to enable, but we we don't really have for organized the way to use it. So sometimes it's just me if I think the future is is either for first camera I, just Market I have one date and good first issue and see if someone picks up, if not, if no one is picking it up after a while just I just do it myself: okay, okay,.

E

C

Are some some of the ticket issues there still Market? That's good first issue. If I remember correct.

D

Yeah there's maybe two okay.

C

D

Yeah, nice yeah and one of them is tracing related. So you.

E

Could take a shot at it? The instruments, instrumentation scope, attributes.

D

Yeah, it looks.

E

Like Williams.

D

Did not do anything about it, um haven't seen the pull request related to it so I'd say you probably.

C

Has to be a science.

D

You can go for it, man, yeah.

C

D

E

E

Sounds good man.

D

You're now assigned awesome, you have expectations on you.

C

Oh I think the service that I'm saying has already been defined.

C

Okay, I'll close this one.

B

So for the rest of us, I think we'll need to First create an issue before we can get a sent something all the goods are already taken now.

D

B

E

D

I mean I, think I think there's plenty of tickets that can be done, I would say, take a look through um the ones that I have Linked In the uh what's it called in the Milestone because, like those are all things that we can at least like start discussing, even if it's, even if it's unclear of like whether it's good or bad, it's like okay, we can at least say like okay. We can do this after.

D

um We can do this before that sort of that sort.

C

Of thing, yeah that we needed.

D

To do it before.

C

Yeah I think we can also take a bit on those issues as if it's good candidate for good first issue or enable them as needed. Yep.

D

B

E

Thanks guys, good plans.

D

um I think a couple action items for this is: uh uh what do we need? What do we want to do? Leighton's gonna work on that one ticket.

D

And then uh yeah I guess take a look at the uh review. The I guess review. The Matrix would be one um and see if there's anything that we're we're missing, uh review, Trio I. Guess we need to triage the tickets. Maybe that could be the next. uh The next meeting, as we just Triage, go through the tickets that haven't been looked at and just like, say like yup, keep it for before. After and.

C

D

uh Or close it, maybe who knows? Maybe it's not an issue. uh Trans is the stable list and.

E

That'll be a Tuesday next week, Tuesdays.

D

Yeah and yeah go forth and do awesome things.

E

Cool any other topics.

C

I think that's it.

E

Awesome: hey it's great.

C

Meeting you guys folks I'm here.

B

Yeah next Tuesday.

A

Bye. Thank you, foreign.