Open Telemetry Uncategorized, 30 Sep 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2022-09-30 meeting

Description

Instrumentation: Messaging

A

A

Hey there, hey Dan foreign.

B

All right folks.

A

Hey Dimitri.

B

Does anyone else usually join this meeting.

C

uh Usually Alex and Tyler do um I've also has been seeing uh Benedict in here lately, but um it might be a few minutes still last time. uh Some people showed up I think about five minutes. After the hour, yep.

D

Very often we don't start until at least five minutes after the hour. Okay,.

D

I have the same problem with all of my other Hotel meetings that people are used to the meeting doesn't start until five. After so now, nobody shows up until five after I I'm, the maintainer of the JavaScript seg and I've been slowly training. People earlier and earlier I start one minute after the hour now.

A

Yeah, that's probably the good guy yeah.

B

Just that right away is actually pretty good.

B

People show up um pretty much in time.

D

Yeah, it's really tough without like at least Tyler, or you need like one of the maintainers here now.

A

That Evan's, a trailer.

D

It's easier, but.

B

Usually around this video.

C

uh These days well, I don't know. um I have been at least the past. Well, it was two weeks ago. I wasn't here last week, uh but previous to that it was usually Tyler or Alex.

A

B

I, don't know they want to start or I can start.

C

um You can start one thing: I wanted to discuss. Real quick was um this issue. I was just wondering if you had any ideas for this um hasn't gotten any discussion for a while. Now I.

A

Can take a look track and went into shows.

A

B

To be honest, it's more like it's not something that we can just decide right away here. uh I I agree we we need to like bought, probably that would be codowners on a regular basis and then that what uh potentially can apply a label like hey, uh consider, considers marking as maintained or something like that.

B

So it's clear like we have been maintenance that many times no one responded and like both suggests, hey this. This uh component clue well, it looks like qualifies for marketing, is maintained.

A

B

This is just my opinion. I can respond to this one after the call, so we can proceed with the discussion. Okay,.

C

Cool attention uh mostly because so the past couple meetings, we've actually gotten back to I, think about the June time frame in terms of triaging issues, I've been keeping up with triaging them as they come in, um so we could start there if you want- or if you had another place, you want to start for this meeting. um That'd be fine too, but I just want to bring that up, because um I think that would help with some of the like the really really old issues, uh just kind of marking. Those is.

B

Yeah we should cover new first and then we're gonna can go back. What do you think? Okay.

A

C

So I can share in that case,.

A

C

um Close so um I got back to it looks like early June, but I didn't note a couple of issues that uh need to be triaged um between uh that point and a little bit later on um anyway. This one was the first one. This looks like a new component request, but I wasn't sure how to classify it. It could just be a needs. Discussion.

C

um I was looking for input there.

A

A

B

Let me can you screw up I, don't I need to see like the so I can find myself. Okay in Harrison inheritable.

A

C

The gist of it is just um propagating uh the attributes from parent spans to child spans. uh It seems that the suggestion is that there should be a processor specifically to do this, but it's well, it's possible that there could be something that already does this pretty easily um I know that the transfer pulled this off, but I don't know.

B

I, don't believe we have anything like this, so essentially they want to propagate the Spanish attributes to parents, no I'm, pretty sure we don't do anything like that, and this probably can go to spawn processor, but at the same time actually no spank processor uh operates on span only so it doesn't know anything about like parents.

B

So this this processor is supposed to know the state and I. Don't know what like the good candidate we have. That knows about the state. Probably like I, don't know like table based sampling. Processor does not does know the state, maybe that guy can do this. But anyway it's pretty like it's pretty complicated task to be honest in area.

A

Let me hear it: that's a good future requirement.

C

B

It sounds like this would make sense as.

C

A new component then.

B

No I I believe this could go to tell them something because or we can move tailbase sample functionality for the state of the trace uh somewhere like in the library and that the new component can reuse that, because essentially, we need to wait for the whole span to like no to the whole spine, but at least to like. But anyway it's actually a cool span, because you one one span can have different patterns right.

B

uh No one's parent can only have only one parent, so we need to wait for one parent spend only, but still we we need to wait for that. I don't know. Maybe a new processor maybe tell best sampling, processor, okay, new processor,.

C

Do you want to follow up with this one then.

B

Yes, I follow up it's pretty big one. Let me let me read it after and I'll I'll respond. Okay, thank you! uh So we can so we can go to the next one. Yeah yep.

C

um This one needs to so this is going to be probably.

A

B

um I, don't even know if we can't do that with like existence, but anyway, let's put It P3 like a lowest priority. It's not a big deal incentive. Sure.

A

B

So yeah and we can keep it- there is no jcd law player here sounds good.

B

Enhancement, yeah.

A

C

C

Okay, so this is from at least a week ago,.

C

B

So they're using the glb receivers. Yes,.

C

Would put it in the core.

B

C

Okay, well I'm not going to tag with anything. I will simply say waiting for author.

B

um Okay, let me I'll pull up back on as well.

A

Okay, thank you.

C

Oh okay, this looks like this is mostly already handled. I, just noticed that it wasn't tagged with a.

D

C

um Yeah those do exist, we haven't been using them as much we've usually been putting them toward a like a specific. This looks like it's the the heck exporter, though,.

B

And it looks like there is uh already some kind of.

C

Yeah, an enhancement yeah.

B

Anyway, it will have a sign in here yeah.

A

Okay shall take care of it.

B

Okay: okay, what was the issue we we discussed like? When? Was the user Arrow likely? Can you please please give me a name? I cannot find.

C

It um that's this istio one Hotel.

B

A

B

A

118.5 yeah.

A

Thank you, yeah we're good or we can proceed one moment, foreign.

A

Okay, um okay,.

C

And give it there says once again.

A

B

D

Is someone else is asking for the the receiver itself to parse the Json yeah.

B

Definitely not the thing we want to do on the championship, it's uh it should. It should be a processor and uh like this is what Dan was talking about like some time ago that we potentially can extract operators that cannot live only in local related receivers tools like another processor. So this is the thing that.

D

uh I think this is probably won't do right. I mean it's handled by the transform processor, the logs transform processor.

B

We don't have local transform processor. Well, all we do have lost transport processors.

A

C

A

B

A

Like, for example,.

A

No, it does not.

B

Waiting for that author right.

C

uh I mean it seems like so Dennis. Oh, this was assigned oh well, I'm gonna, sign him: okay, yeah! That's fine, we'll just saying wait for author I think that makes sense. Okay,.

B

We can have another rule to automatically remove issues that are waiting for the author for more than a month. That would be a good idea.

A

uh Just appear forward.

B

Signed so yeah we have a signee. We have labels, probably new unit priority I. Believe priority is low here.

A

Looks like that's.

C

Yeah, it doesn't seem very high, yeah. Okay, um this one just needs a priority, which is probably just going to be.

C

Oh, wait was this already.

C

This one seems fixed.

A

C

Let's go up here, just wrong record fix.

A

Being called for the wrong metric yeah, it looks like it's fixed, okay,.

B

um You can close it and if it's mentioned that it's not the case, the blue, zero yep.

C

Cool best kind.

A

Yeah data of The Collector data, yeah I, like this that's awesome.

A

B

Yeah wait yeah, there is like bigger Sask and we don't want to do it anything like that. I mean we have some functionality that potentially can uh can be good for them.

C

Okay, um okay, so that's um that puts us at here which these are all fairly straightforward. It's just a matter of um applying it there's a lot of these.

A

Oh, oh, but back to Sentry, okay,.

C

uh He did a pretty thorough um audit of everything. um Okay, I'll just apply.

B

C

Those are all straightforward.

B

Okay cool, so what do you want to apply? Let's play.

C

uh Oh no I'm, just saying I'll apply the the label for the component, so it'll.

A

C

Can figure out what they want to do with that? um Let's see, but next um this one I can just apply a thing to this needs a proper label. Okay, that never got.

A

um Exporter in blocks.

C

uh This sounds like a question.

A

Yeah, so let's try it.

A

C

All right, let's look at this one.

A

Oh, let's go up, uh it was a battle.

C

The title is: defining a tailbase sampling policies.

B

This is a significant task. uh Let's see what the addresses said.

C

uh Okay, I'm gonna, say waiting for author.

B

And we should assign that there will be something here. They will it's the tail, but still with something processor.

A

B

Tell sampling okay, okay, cool.

A

A

No critical missing components: um I can't tell.

C

I'm not clear if this has been.

B

Yeah, it looks like it's colored.

A

A

I'll just ask yeah sounds good. Thank you.

C

uh If this is still open, it might be worth re-tagging these actually now that I'm, seeing these.

C

You might want to paint the code owners again to confirm um same with this one and.

A

C

Yes, it still seems active, um and this is the eager receiver.

A

What was this okay, quite a while ago.

B

I decided to them if there is already an apple request.

C

uh No, they just said they're seeing same no, they just mentioned it. It wasn't necessarily being resolved. Okay,.

A

C

Think we should just retag it and let the.

A

Okay, this looks like this one's.

C

Actually, being handled.

B

Yeah, so we're gonna assign all right. It's already the side, cool.

C

Yes, I'll just tag it, so it's.

B

It's in your company detection; no, no! It's a new component! Oh.

C

Is this a new we've got a greater receiver I see okay, so we want to reapplied.

B

Yeah I believe we will get a new remote writer remote right, export directly.

A

C

One I'll just let them they clearly are working on it: okay, uh blocked by that we should avoid them.

C

Well that's been closed. So.

B

It's a Minnesota detection, processor.

A

Okay, thank you.

A

B

Still Google, like part, the sort of detection processor and the detectors Google detectors, so both I need. It.

C

And the Google yeah.

B

Yeah Google is needed as well.

A

B

Right and we probably need to being.

A

C

We just ping him and ask what the.

A

Status is yeah.

C

C

Okay, so it seems like this is fixed.

A

Cool yeah we're gonna close.

A

What are we doing on time? Okay, all.

C

A

Probably just tag that real quick, it's probably need to be.

A

A

C

Okay, so it seems like that person is oh okay,.

C

I'll just reassign it or.

A

That's all right here: it's uh it's in progress.

C

Okay, so these are still.

A

C

Oh three days ago, okay, cool.

A

C

This seems like a question.

B

My kind of suggestion for.

A

Improvement I believe so.

B

Anyway, we can also add the load balancing exporter.

C

Wait is there usually mean to load balances.

A

C

There is wait, load balancer is made with penis looks hubs. Oh I see, okay,.

C

Sort of DNS to create hash rooms.

C

Maybe this oh, this seems like actually an enhancement now that I'm rereading it.

A

D

I guess the current one just uses basic DNS load, balancing.

A

um Okay, cool cool: we should be good, we are on.

C

The real time, all right, we're at time. Okay, thank you. Everyone thank.

A

You folks, all right thanks.