Open Telemetry Uncategorized, 22 Mar 2023

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2023-03-22 meeting

Description

Open Telemetry Meeting 1's Personal Meeting Room

A

Hey hi Mark I.

B

A

It, let's probably wait for a couple of minutes: I think Sr is not going to join today and he just paint and uh couldn't wait for Tom. Let me see if he's around okay.

A

C

A

S: okay, let me just share my just a minute.

A

If I screen is visible, right, I mean I've just shared my screen. Yes,.

B

We can see it yeah, okay,.

A

Okay, here, I have actually I mean probably I'll, be here for another 30 minutes. I have some some personal I think uh thing to release sort out, I think after 30 minutes so I'll be leaving after 30 minutes. Probably you guys can continue after that.

A

Feel free to continue so okay, specs, I, haven't gone through. Anything probably will have more consolidation unless, until you have something to discuss, probably just let me know otherwise, I think we can move to the uh next one I mean to the PRS.

B

No I haven't noticed anything in respects.

A

Yeah I haven't I've, been checking the mail, but I didn't see anything which probably needs uh attention immediately, but yeah, probably I. Think next week will have more thorough.

C

A

C

This morning we have one topic: it's I think. Yesterday we had a meeting right at Montana's meeting. Yes,.

A

So I think that's already there. Okay.

A

Here, Tom, you want to talk about this I think we already described. I can just shortly say probably I think we had a discussion uh Mark uh with we. We I mean me asan and Tom yesterday in general, uh to add you as a maintainer. It was not not specifically discussion or not specifically whether you are the right choice or not. It was whether having too many maintainers and will will disturb the balance or not so I.

A

Think in general, we all three agreed uh I mean as per your contributions and the way you have been working in the project, so I think we all agree that it's, it's definitely good for the project to pull you as a maintainer, so I think there was no concern in that it was I. Think one.

A

The only concern was that if we are going to have all of us as maintainers probably have a proper process laid out in terms of a project charging and all those things some of the things I think there were some concerns was raised by Assan and also and I.

A

Think, probably me also, and so asan has some idea like he want to create some project and have proper trading so that we all agree what all uh what what what all work we are going to do with the index couple of months in terms of the issues prioritization centralizing. So you want to create some project uh uh on in that line and uh I think probably is some time frame was next couple of weeks or few weeks from now and then I think. Once we have a proper process layer not laid out uh I.

A

Think we'll we'll just add you as a maintainer, so sometime next month, I think we should be good to add you as a maintainer, so I don't see any issue in that or Tom. Do you want to add I think what all probably one one thing was to have a proper process in terms of hygiene and prioritizations of the issues which we are going to work so as all the maintainers we all we all are in sync that what what next is going to happen in next couple of weeks.

C

A

But anything else was there or.

C

um I think that covers the most I think one um a USM will create great work items right, yeah.

A

At least two work items, one one was for project rising and proprietization so that we all uh as a maintainers are in sync with what what's going to happen in say in the project in next next month and two, so that so that there is no conflict among all of us and we all agreed at what what what's?

A

How in what direction the project is moving. So that was something I. Think the uh I'm going to create one issue for that to have a project, at least at least in and have a proper trading in the issues so that we all so there is no conflict in terms among all the maintenance and anything else was there. Apart on that, I'll create one issue for that and I think asan said that he will. He has some experience working in in nginx uh on that and probably he'll he'll.

A

He can drive that, but apart from that, anything else was there that we discussed, if not I, think probably this should be fine.

C

B

So this is prematurely about planning to see how we plan things and yeah.

A

Yeah so and I think it's probably the initial initial uh set of set of uh process to be laid out and then probably as a as we all as a maintainer can always enhance that and make it better. So it's.

A

So that that was something I think probably.

C

We agreed in terms of.

A

It was just that uh went too many too many cooks can functions, spoiler recipe. The recipe is not proper, so let's have a proper process laid out and then probably we all can start working together as a maintainer.

B

A

Or in the next few weeks, I don't want to have any deadline, I think I, don't know how much time asan will take to drive this and finish it.

A

uh Would that be fine, I mean Mark or any you have any concerns in terms of, uh if you want I mean if that that works, fine, for you.

B

Well, it's a bit unclear who will uh participate in that process. You see uh the current containers which needs yes,.

B

A

So that was, it was said that, as of now I think the current management maintenance for the initial process, as I said that, once once we pull you in, we will always make it better and enhance it. So it's uh I mean based on the agreement of all four of us, so that was just the initial process to ensure that once we all four are as a maintainers, they should not be any conflict in terms of what we are going to work.

A

uh So so that's that was the reason why we were delaying for a few weeks before pulling you in as a maintainer.

B

A

um Again, it's not for you, even if there was even if owent was there I think it would have been the same thing. It was just that we don't. We wanted uh to have a process in in place before before. We all four work together as maintainer, so.

A

I mean that was a few. That was a thought.

A

B

Okay, well, let's see how it goes. Yeah.

A

Yeah, let's see probably I, think there are still some gray areas and I think hopefully, hopefully I think we'll work together at least I'll. We will hate automatic. We can probably have one more meeting in between to see how the process is going because I don't want this I mean at least we don't want this to this should get delayed of of so probably I think that, let's see sometimes early next week, if you can pull Mark as a maintainer anytime I think that should be good. Okay,.

C

That should work.

A

Yeah, probably we can have one more meeting in between uh maybe next or uh early next week, early next month to see where we are in the plan and and then we can I think I do in terms of permission, I think I, don't see any issue in giving the permissions, for uh whatever permission we have a maintainer is have I think we can all we can provide it. So uh foreign.

A

A

I mean this should not stop anything which, probably you feel should be done. If, even if it's a release, uh doing a release or I mean I, think it should be even the merge permission for merging also should be fine. I, don't see any concern in that also.

B

Yeah well just just remember, to keep a map queue. For example,.

A

Yeah and yeah I think one one process was one in terms of project and second was in terms of the uh the merge process. I think when, when we are going to merge in terms of the number of approve, how many approvals we need need based on the size of the pr, so I think that was another thing.

B

Well, I'm surprised that this thing should have been defined by now, yeah.

A

It is already defined, but I think probably we still sell I think there are. Sometimes there are some some confusions based on the size of the pr I mean we still have some some confusions I mean when exactly we should be doing it. So I think it's merge. Policy policy is not still very clear in terms of the uh how complex the pr is.

A

uh I mean if it's a small PR, probably we go with our Instinct with a small PR I think just one approval is okay as of now, and we just do it, but it's not something document will documented so.

A

It this is just to ensure that there is no conflict among us as a maintainers when people are working together, it's not something again uh um as an ear to a particular individual. Okay.

A

A

A

Okay, I think, probably, if anything else in this uh and adding timeline as I said that probably let's see early next week, if we can, we can complete this, at least in terms of triaging project prioritization and have a more clear, much policy. I. Think probably we can just pull you so here.

A

Okay, I think probably we can quickly go to the other, PRS and sure.

A

I haven't seen this ignore more warnings of generated photograph files in general. It should be fine to ignore those warnings, even though they are not included in one and extra Mark. Thanks for reviewing it.

B

It's a small cleanup, promo event, but then.

A

A

Yeah, that's probably copy copy based things.

B

A

Okay, yeah I think probably interviewed.

B

He's not online now so by the time you see that.

A

A

I have created this Implement Force flush for periodic metric reader. uh I think this wasn't something we would as of another. Force plus was not doing the actual uh collection of the metric data and we were doing because this was a periodic reader, so it was doing collection uh every configured number of time and then it was trying to export uh one last collection. A wind export was called, but I think this was not the right logic to do it. So I've removed the collection before I mean just when they said shut down.

A

It called, but instead of that uh as part of force, plus we The Collection can happen. uh Collection and Export can have happen, so I have done the changes, but somehow I think it's bit complicated, I realize now uh and probably it needs more Cleanup in Sports flesh functionality for not just for reader, but also for trace and logs.

A

uh So I think that this is a bit complicating it uh um I mean the reason why it is complicating is, as of now I think that if some otlp exporter, even if it's getting the time, the timeout as zero still it is, it assumes it uh to be infinite time and it uh so I think there is some some special handling, which is there in the exporter code, because of which, when we are calling the exporter Force flush from the reader, Force plus I, think we have to set the time accordingly.

A

So, yes, probably this needs some Cleanup in general. uh For what uh timeout value should mean timeout of zero should not be infinite timeout. It should be really zero, so yeah. That type probably has to be handled.

B

I went to some functions, yes,.

A

And I think that's a valid concern. I didn't realize it, but now, when I was implementing, this I could see that that is causing the logic inside this code to be a bit complicated and probably it has to be simplified. Okay, remove unused header file, I, don't see any progress in this and.

B

I think it would not be.

A

Easy thing to do it.

B

A

B

One I mean this guy uh started as a very good initiative to do some cleanup. But then, if you look at the the pr itself, I mean he's touching more and more code and um he's doing Force push so that every every committee is replaced by the next one, which is and of course the whole thing doesn't build and and whatnot. So, let's see if he comes back and if.

A

B

Have some questions about um we? We should do that in general. Yes doing that with 200 files changed at once from by someone who does not move a code base is going to be, they see yeah.

A

And I think it should be done in in component wise I think, instead of doing everything at once. Oh.

B

Yeah piece by piece: yes,.

A

Well, I think probably he realized it's taking lots of time to do it. It's not easy thing and he's not coming back after that, but yeah. Let's wait, I think that's.

B

A

Yeah, okay, uh yeah: this is on this.

B

One we wait. Yes,.

A

This we already discussed yes.

B

A

Is now I think unblocked right? Yes,.

B

So real tools did already with a bug fix there that allows to generate code again, so I could update the semantic conventions properly. Then.

A

Okay, I think this I'm just putting it okay to merge. Probably we can give it a time today if somebody want to review it. Otherwise you can should be good to merge. It.

A

Yeah I need to review this. It's on me and probably I'll, take some time to review this during this week.

A

B

A

No, this is not a breaking change. It's a fix of.

B

It's it's real, drawing the transitive dependencies between libraries so that it's all.

A

B

From the same place, yeah.

A

Which simplify.

B

Things forward, income.

C

A

So it is just ensuring that uh this Library will cont will surely contain all the symbols from grpc uh from libgrpc and sometimes in some scenarios. Those symbols are getting not copied to this library and which is causing a problem. So this this is handling that scenario.

A

So I just wanted to test it, uh because this this this setup is I'm using all the time, the same setup for which the problem is coming in some somehow I never get this issue, so I just wanted to test it before um yeah before approving it in general, the the person who created this issue I think he has already confirmed that it works fine for him, so yeah and.

A

I think we know nothing on this right. Api.

B

Check, no, it's uh well. It's work in progress. First of all and.

A

B

We can discuss that later. Yeah.

A

That's fine uh enabling build I have to close this sorry I'll do that here. This is on with asan. Probably he can take it. This also I need to close it. Sorry about that and I think these all are still in discussion for review.

A

And I think this guy is not yet okay, probably I, think you can think him once.

A

Probably, let's refresh it.

A

Okay and same thing for this.

A

Yeah I think Tom. Probably you wanted to look into this right whenever you have time.

C

Yes, yeah, look yeah.

B

So I haven't I, haven't looked at the details itself for this one, but I'm surprised that it takes 23 commits to change free files. So I, don't know why this is not converging and.

A

uh I'm not sure because I mean I I saw this and probably from the last time when I reviewed this I think the changes looks good to me, but I think there were some valid concerns by Tom Tom. Do you want to talk about this? If you have, if you still remember this.

C

A

I know it's being some time back, probably.

C

Yeah I think the basic package we set the variable, but we need included the file the torture file. Man explicitly in the code in the past, currently I think he's hating is for g-test with the way I think we didn't do that. We didn't do the include so I think that could be the error because for the quarterback or that pass we we did the includes, and then that works it works there.

C

That's why we said that there of the semic variable for the iron scope, because we just said that when there's there's no, it will not make any difference. We need to include that blockchain file explicitly.

A

C

And that include is currently missing. I think that that's the issue.

A

Okay, but I think he's saying that even after doing this include is.

A

Okay, have a look into this term. I. Think probably this is this is something which is spending for long, probably I. Think, let's try to close it if it's possible yeah.

B

Okay surprises like if you look like it's 22 commits for 12 lines,.

C

A

B

A

Yeah I think this is not we're going to do anything on this, and this is I need to review it. Mark I think. Let me spend some time on both of these two this week. uh The good thing is that event has already reviewed it, so that gives a uh good confidence on the changes.

B

Yes and so well, it's uh it's work that I started a long time ago, so it has uh evolved uh along the way, but now it is stable.

A

B

The thing is, it's uh I also depend on that, and so I really would like to uh for this thing to to move forward so that it can.

A

Be moved at some point, you're feeling, okay, I think that's probably spend some time to review this this week.

A

I think that let's try to see if we can merge it- maybe not this week, probably by end of next week, if we can Target to merge this uh I think I just want to ensure that we review it thoroughly before really pushing it. It's a big big PR! That's the reason, but I think I'll spend some time this week and next next week to ensure that review it thoroughly and then probably, if it's all good, we can merge it and include it in the next release.

A

uh Okay quickly, I think I have just four minutes. uh So probably, let's see if there's anything new um uh get span. Duration. While it is running this is not a valid I mean this is not. We don't support it. It's not as part part of specs. Also so probably let's see uh and ignore more warnings, yeah I think event is working on that this I have a PR I. Just wanted some comments, so I'll just probably resolve those comments.

A

Yeah, this I think Mark. You have created.

B

Okay, well I I created that when the other guy started to do some Cleanup in Edo files or.

A

Something yeah, yeah I, think that would this would be good to have include what you use build I haven't used it, but probably I think it will help us to have a cleaner includes.

B

And it well, it will help us with dependencies in general, yeah.

B

But not nothing urgent anyway. It's a small items, I mean it's a.

A

B

Task that can depicted by anyone yeah.

A

Exactly I think this we can close it uh yeah. Let me see I think probably, if you can close this.

A

Yeah this is mentioned is help wanted because we it's not mandatory. It's optional as part of specs I created some of these and probably created by Tom, also so yeah and yeah I. Think, probably in the interest of time. Probably we can move to if there is.

A

uh This should be close right.

B

Let's see so that was first I think on on Mac OS yeah.

A

B

Think you did it.

A

It's a model, and that seems to fix it yeah. This is this- should be fixed here. I think: oh, not this! Okay! Let me just check I, think probably I. Remember we fix this issue and uh it should be fine.

A

Yeah I'll just check that and okay, probably I, think in just in the interest of time. We have time lock, API, SDK, Beta, release, I think we have most of the things uh in place for logs and the only only open issue right now was for the namespace.

A

Let's see if there's anything yeah this was this was the one of the open issue if we should rename it uh at least rename the namespace to logs Bridge or something instead of keeping it as a logs, so probably at the API level. So let me check. Let me probably comment it here and check with oven.

A

Tv has any concerns, if not I can or Tom, or anyone can pick this up, and we can just finish it up and then uh I think we should be good to have a one Beta release and then, let's see uh the plan up for the actual 1.0 release afterwards.

A

But you I mean with beta I, think we this means that we should be able to mark it as a beta status as a beta, at least in the documentation, and make it as a Beta release here, instead of experimental.

A

And yeah I think and that's one quick thing was using stack Overflow instead of GitHub discussion. This was something which came up in the last maintenance meeting.

A

uh It was more of a recommendation to all the maintainers that, let's not, let's not promote the use of GitHub discussion, but let's ensure that any questions which are coming in GitHub discussion either we can, we should be moving it to stack or throw, or we should ask the consent person to raise that in stack Overflow and let's answer it- there uh stack Overflow is just easily searchable and indexable, so I think it comes in the novel search um and more more benefit. People would be benefited with that. So.

B

The stack Overflow is highly visible, searchable and so on uh also from actual experience where it's very toxic, you know a lot of people who are just shutting down every idea. Every minute it is typed, so I'm not sure it will, if we can actually work with stock over food.

A

No I mean why toxic I think so probably just come again is saying: it's well.

B

When someone asks a question, the question is shut down immediately close or people sometimes.

C

B

Case out of 10, someone will help, but otherwise people are not very helpful. In general.

A

Yeah but probably I think it's it's something we as the corner, as at least uh contributors to C, plus plus. If we ensure that we are going to answer them sure those questions there I think that it's just that our uh instead of spending time on GitHub discussion to answer that we can, if we can answer it on, stagger flow or even that's totally fine like if we answer it on GitHub discussion.

A

But if we can even put that question on strike overflow and put as an answer also I mean it was just promoting uh stack overflow in general for open Telemetry for.

B

All the six well, what we can do is have an open, Telemetry tag there and start answering questions. So what what will work uh but be careful of other people also commenting there and maybe not providing the proper the correct answer? What can happen too.

C

Yeah and allowed we're past 30.

A

Okay, I can just leave, and oh.

C

Okay, I'm fine yeah for this issue. Maybe we can discuss this in a with in the issue right.

A

Which one the stack will flow well.

C

A

Create an issue- and probably we can, let's not discuss it in GitHub discussion, but we can create a issue yeah, okay, yeah. Thank you thanks for joining and it's probably we can meet next week. Yeah.

C

Thank you, yeah.

A

Thank you, bye.