Open Telemetry Uncategorized, 6 Apr 2023

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2023-04-06 meeting

Description

Instrumentation: Messaging

A

B

Hello I think that's probably a million Johannes are late. uh This time.

B

Okay, yeah I called you your first hey and after that yeah, no more, no more sound from your side.

C

um Does it work now I guess so.

B

C

Oh yeah, so Johanna said that he's not joining today he's on vacation.

B

Yeah I mean remember now.

C

Yes, so I don't know if uh other folks will join. Let's see.

B

Yeah totally I mean they put some items in the agenda.

C

B

Yeah yeah I mean he was iterating on his uh Otep.

A

B

A

Then yeah probably.

B

The interesting thing other than his own iteration is the the issue he posted well actually Ludmila posted regarding Cube versus topic um yeah. She posted an answer um which was interesting.

C

uh I guess I'll start sharing my screen then perfect.

B

Yeah that works.

C

Just just a sec here we go.

C

All right, so what you're talking about is this issue here right about the yeah, so I can tell that I follow up, followed up here internally, with some folks I'm still doing some more uh digging, but so far I didn't find anything. That is really uh this. This that this attribute kind is really um doing anything special.

C

um We also record it. We also collected it in our instrumentations uh I think it's basically the same name or it was destination kind and Source kind, with the same values, q and topic, but so far what I learned is that, apart from yeah, basically what um what Christian said here apart from showing an icon or offering to group things, uh not not really much but I I'm still too, to find more with some I have some meetings with some folks too, to ask.

B

C

B

Yeah, that would be great yeah, totally yeah.

C

So we'll see, uh but it doesn't look very exciting uh at the moment, um so it might be just something that we can make it optional. I think as Milo mentioned here, uh make it optional and and keep the values that they were before Topic in queue. And then, if there are more that can be added.

A

C

Or we don't even do anything, because what I think she mentions is that some systems have way different things like rabbit. They have exchange type, uh so I'm, not sure if she proposed or her intention was to um add exchange type to the list of values or an entire different attribute called exchange type.

C

uh Because that's the case, then that doesn't make sense. um Yeah.

B

Yeah yeah yeah yeah, okay, uh good to know in that case,.

B

Yeah, so my feeling about this one is based on what Dwayne mentioned. My impression is that this should be not a requirement for the first version.

B

um Honestly, um so I could say that if we can come to an agreement, it's great, if not that doesn't mean we won't have that.

C

Yeah yeah and then apparently there is more things as well to capture like little mirror set here. There is a settlement kind, rich, but I think these are additional, not direct, directly related to yeah.

B

C

This um yeah I I, don't know I I'm not opposed to any any other other things uh I think if there is books either like he has apparently Dwayne has um value or wants to have it uh I, don't see a problem in in rolling back and coming with it again, leaving it uh with Topic in queue. I think the also think that was in the that sparked it was that we should open up the enum, because I think it was like this.

B

C

uh Nobody was that uh allow additional values or something that was yeah false and then it was proposed to open up to true. So it would basically put anything and I think that that when the discussion started so yeah.

B

I, don't know if you have court, but now uh with some proposed changes, all the Norms will be open.

C

Oh yeah I think I saw that yes yeah yeah and then then basically uh that basically solves it. I think yeah yeah, because if we can just comment con with this back and then if there's people want to say P2P or Pub sub, because I think in some other messages that makes more sense like Pulsar I, think yeah they can. They can add that um yeah, but I still think. If it's something that is well known, it should be in the semantic conventions. um Yeah.

B

Anyway, yeah, let's hope that Dwayne provides feedback, as I said before. I think that this is not bad to have, but given that there are other priorities, I could focus on those ones and if Dwayne or anybody else can provide it's almost wrong argument, then for sure let's actually include this. Otherwise, it's an optional thing and, as I said before, we don't if we don't prove any value right now, we can just wait for the second version or even though, as a follow-up, you know yeah.

C

Yeah, it's not not if we, even if we add now it's going to be optional anyway, because I think some some in some depending on the library, you don't you don't there's no way to to know this um uh for like from the producer or consumer. There is no way to know so it has to be optional and then, if it's optional, then it can be added later on. I guess: no! We don't no problem at all. I think yeah.

B

C

It makes sense okay, so let's see what what things uh move from that? What do you understand as always, to look at the board? So maybe we can do that quickly, not.

B

Sure, if there's.

C

Any movement, but we can do it.

B

C

Yeah, um so let's see I will start from bottom to top, so yeah I think this is uh blood. Miller, create this issue, but I don't think there was anything yes, um I think this is they want to. They want to discuss this this set out during the meetings here. So um there was no movement in this one.

C

uh Yes, this was something that Johannes wants to do. I think I'll I will offer uh if he wants help on splitting things, uh because this is a no-brainer I could I could do it on my spare time.

B

C

Guess uh between tasks, so he wants to uh we discussed and we want to create separate emo files for the different systems instead of having them in the same uh generic or general document yeah. But no, no movement on this as well.

C

Clarify semantic um yeah, I think this is also no movement. It was something about the spiral a long time ago. I think he has a uh the table that we have there with uh the spam kind. It's a bit confusing in some weird cases um yeah. So this is something that also we need to take a look.

C

I think this is going to be a bit complicated to be honest, but.

B

Yeah, actually I am wondering whether these will change only the interpretation for messaging or we will end up impacting the entire. You know, meaning of a.

C

Standpoint, yeah I think this is the second to be honest, but it was a long time ago. I don't remember fully, but I think there was some um like profound uh problems with what we have today in no no way to express what we have, what we have there without being super confusing um uh yeah, but I guess. This is also something that we need to discuss. As a group um Amir was was talking about it a lot but uh I'm, not sure. If it's.

C

If it's going to be uh coming um to to the meetings again, maybe I.

B

Think that it's yeah this week is Easter, um so I think that some people are probably thinking this.

C

Week- yes, probably yes, uh yes, this uh other one here, I think there there I will link this to something else, um because in the span, structure and I find a comment here in the Otep. That Johan is writing there was this coming from Christian as well.

C

Apparently there was some discussion in the dotnet repo about the fermentation for AWS, Lambda sqs, and what what Christian is described here is that in a batch in some cases this was in the concept context for for AWS for I, think uh SNS, even though the system but anyway, so it was like when you send a bench, you could um use the last message on a badge to edit at that at the context from the last message in the batch as a parent.

C

um uh In the the case there, that's it's a long discussion. This link that he posted here- I read it today, but it's a basically for that particular service in Amazon, Amazon SNS, you send a batch, but it's just the API is like that. So it's always a list, but is is always just one message so uh for that case it makes sense, but he discussed here that maybe we should also put a diagram or something that highlights this case and I.

C

Don't don't really think it's a good idea uh to be in the generic uh uh specification because I think it in most cases, that would be really bad uh if you use, for example, the last message in a batch to say it's a parent because you could potentially just um mix the traits together and I, don't know, put the processing processing of one message as a parent of another I think that will be really messed up.

C

Yeah, that's basically what they're discussing in in this year. So. But the point was that in both of these things it was uh brought it up the uh uh not or swap the link or use instead of adding a link at a parent- and there was already this issue here so I will I will link them together later, because um I think the discussions are are sparse. So now, there's discussion in here discussion in there and I think I probably should centralize them in this.

C

One and then we can capture capture in our board um I'll. Do that I'll. Do that afterwards.

B

Yeah that sounds good, I, think yeah, yeah I also don't think it's a good idea to have this in the generic portion of the spec yeah.

C

B

And we we will later get complaints from other messaging systems that don't work well and yeah.

C

Okay, there's one from cause of, inter designed to me: I I will wait until we are closer to to having this Otep merged and then I will take a look at what the impact for this is um probably have to change the existing conventions that we have today, but I, don't think it would be too much of too much of a problem.

C

um uh Yes to add example, batch and I think this. Maybe we already have this yeah I think this also probably needs to be linked to the old type. Well, I'll take a look because I think we already have examples on batch and with the diagrams yeah.

B

C

This one open as well.

B

Yeah exactly like there in the tab itself, yeah yeah.

C

Not sure why it's still not linked but I'll, do it um yeah this one is just a one small one. Maybe I will take a look if I can take this one, because I was also looking at internally. We also record this size, but this also might be that uh it conflicts or it will come as a solution with the ECS proposal.

C

Now that was merged, um yeah I have to look that this is uh because there is a bunch of these attributes in HTTP and they differ from messaging like payload size and compressed size and and then there's math and then yeah this just to to make sure that but yeah. If we have this in HTTP, then I guess we can just follow the same that they have probably not.

A

B

C

Just reuse yeah, but they will be in HTTP, so yeah we're using, maybe not possible, but yeah.

B

That's interesting: you would expect that there's more value from for from those ones in metrics, but okay.

C

uh Yeah, no I think I think um yeah for for messaging I think he mixes, because there might be limits for the message size uh in the vendors, for example, I think.

B

um ah Okay, I see right: okay, yep yeah.

C

Like in Amazon or Azure, maybe there is a limit for the size of the message or you pay for it for it, depending on the size stuff like that,.

B

Okay, that's a good one. Yeah makes sense.

C

uh Yeah, the the ECS one I think through the mirrors, is on it. um I think she's. On top of all these, yes I thought the the proposed that she did with the HTTP, so I think that's just moving uh yeah. The Matrix is I'll shoot purple or she demoed uh uh last week. This one I'm not sure about the metrics. What what we will propose, but I.

B

Think, oh I think you were not here last week, correct or wait. No I was here. I was serious, okay, so yeah. Basically she proposed something, but you know, like I mean she presented with she thought yeah.

B

A good set of metrics I. Think that that's all for now, and as far as I understood this will be Marcus experimental. You know the Matrix portion will be experimental, but she still wanted to exercise them. You know.

C

Yeah I think yeah I, don't know if I think. The final idea is that the metrics, like I, think the initial idea was that before declaring the semantic conventions table, we needed Matrix as well. So like a combination uh but I'm, not sure anymore. If that's still the case, if the, if Johannes had in mind that to go with the trace conventions, we also need the metrics convention to be declared the declared stable yeah, uh but I, don't think they have to go together to be stable right. The we can just declare this. That's.

B

Right yeah I mean yeah. We can ask. Let me level my impression is that she only wanted to exercise them, Mark the messaging trading convention, stable first and at the same time, experimental um yeah. So that's the plan um just for your information. There was some comments somewhere by Josh sured, um and he was mentioning that he's.

B

um There's always a reason that if we publish something even if it's Marcus experimental people may people may start using that all over the place right which will force us to uh to Market, is to Market as de facto stable or something like that. Otherwise, even though it's experimental will break everybody, so uh there's something to keep in mind, um so we may even we.

A

B

Even not you know when we announced that this is stable, just even not mentioned the Matrix part or so I. Don't know, I mean that's something that Josh sorry will be thinking about. What the adults use for your information. Okay,.

C

Yeah because I think the current convenience that we have are already the de facto because there's a lot of people using it I mean not a lot. I, don't know a lot, but there's a lot of already libraries implementing them. I think.

C

You just wrote this time, so I don't forget. We want to.

A

C

Okay, we'll see what they say. um Yes, we can then I guess um about the links of the creation. Do you know if anything uh moved on that on those things.

B

That suddenly hasn't moved uh so Dave Young, who is in the GC he was trying to help. They go see to find a way to make this happen without breaking many things. um There's I, don't know what to do on this front. Basically also Joshua, McDonald I was proposing that that links our events in indeed and uh basically attached to a you know the way we represent Links at this moment. That would say the same, but we would be attaching event information so.

A

B

Could be kind of hybrid uh and I think that discussion has to be solved even before this yeah? uh So the good thing is that we have time. The good thing is that people I mean the person. The first person opposing Josh idea is uh Pokemon, so I'm hoping that, because they are both in the DC, this can be solved relatively fast. You know, like the discussion takes place fast and basically, a hard call is made later on. Once that happens, we are unblocked uh from this one right, yeah in the meantime, I will get.

B

uh I will try to ask that what's the status with the go seek because I think that one of the main things is that you know we don't want to I mean I mean we were trying to find that work around uh because of them and.

A

B

Don't want to yeah, it's like you know, yeah anyway. So, let's see what happens in that regard. um The other thing is that I was actually wondering whether it there is any messaging system that is extensively used from go. You know, most of the story can see in Java. Is python JavaScript, oh I, don't know I, guess I guess they do. But how common is that like like? They would need to add it right away? Can they yeah I, don't know, but.

A

uh Anyway, long.

B

Story short, we have to wait. One more issue: uh I can actually I will link the issue to this uh original issue on.

C

That one yep okay yeah it's it's not a, uh unfortunately, not that easy thing again. So yeah.

B

But once HTTP semantic conventions is marked as stable, we will we will Define in this group deadlines and then we will have to start pushing people for real yeah. So we make progress so that time will will come eventually. So we should start thinking about what to do. I think that all of these blockers, all of these hard rocks, should be removed by them. So.

C

Yeah yeah, then I think things will progress uh smoothly, probably yeah. Okay, perfect, then I think what Johannes wanted is to go over the comments to open I. Guess we can do that quickly and then call it a day, I think um yeah.

C

uh So what I I was looking at actually uh uh on it today, let's go over there. So this first open comment was what I mentioned so uh yeah he's already. He already answered it, but uh yeah.

C

uh It's all about this, this situation of using the last message and other things, but uh yeah we'll see what I don't think nothing will come to from fridge and fruition of this but yeah um yeah, and then there was some comments from ludibila, but I think this was also I. Think we discussed this in last week and and he she just needed to write them down and then Johannes will fix it.

C

So I think he is already aware of those um and then what I, what I did today is I was looking at the at the document again and I found it actually I just noticed now, but I found that um isn't it weird that the arrow here goes from published to from producer to Consumer on the link shouldn't certainly be the opposite way.

B

I think, logically speaking, that makes sense, but when you think of how a user will see this, you can create confusion so yeah we can probably we should change it actually on the initial Dimension, where jmic dimensions that um that actually links should include event, information I think he uses the arrows in a different in the opposite direction. You know.

C

From from consumer to producer right.

B

C

Because the consumer is linking to the produce, not the producer linking to the consumer, so.

B

That's correct, that's actually he put it those because he says that they can be bi-directional. So basically,.

A

C

B

In this case, in this example, it could be from consumer.

C

To producer yeah and I I I was curious about it. I took a look and I think it's not I, don't think you know I'm certain now that it's not possible with this mermaid uh thing. They don't offer offer backwards um backwards, uh Arrow or link I think they call it link. um I I found this.

A

C

It's 41 already asking for it, but you can't do you can do this today like forward and you can go back, but you have to to swap the so in our case you would have to do. Where is it? uh You would have to put the consumer on the left and then that's what not? What we wanted.

B

Yeah exactly yeah great, similar problem, yeah.

C

Yeah, so that's what people are complaining here: it's like okay, I can't do it, but I don't want to invert my diagram, because my diagram makes sense that way- uh and there was this- this user here that that said, I do this and it works uh some some huge hack, uh yeah uh and then I tried it. It actually works uh with this code. Here it works. So then it will look like this um right, um so I I commented yesterday.

C

I think I think in some of the examples we need to change like these ones. We need to change the order to to be like that.

B

C

All of them, I think, would have to change, because all of them are linked like this and then the other thing that I was thinking is that he has example here this one here, so the one on top is only for a link, so it only demonstrates having a link, but the one below also represents that um you can, you can add, as a parent, if you want, if it's a single message and then to me this doesn't. Doesn't it's not clear to me that this is apparent, um not sure.

B

C

A

Yeah I remember that.

B

Comment: yeah yeah, you made I remember seeing that as well. Yeah I mean it, it won't occur just making. You know just how he wrote link just great parent. You know exactly.

C

um Yeah and then also I, think for this specific case, I think it makes more sense to be top to bottom uh diagram because he has like now left to right like this. He has left to right now and then I think for apparently I think if you change it to this top to bottom thing here, uh then it becomes like this and I think it makes more sense because you can see that um I don't know.

C

This is how you usually consume a trace and- and it's not like uh one after the other, but it's at least below yeah, so I I wrote this suggest. Let's see, um yeah.

B

C

It would be quite some work to change then so I offered a few once I can send a PR to his PR.

B

um Yeah I think that he would appreciate that uh I mean you are fine doing this like by hand, you know.

C

Now that it took me a morning, then I think 20 would not be a big deal. So all right, I'll see I'll text him once like. If he's okay with me, sending a PR, then I'll do it.

C

um uh Okay, yes and then the other thing that I saw that I commented for the meeting was: has a diagram for so for pool based when the application pulls the messages. He has one diagram this one here uh that talks about like it is the liver part of a bet of message to Consumer process process.

C

Stance for a single message can be created but are not covered by the conventions which is, which is fine, um so I think it's missing a diagram like this for the push based, because here there is nothing uh uh there's this, but this is it's for patch I, think yeah, so I I wrote that that I think it makes sense, because it makes sense that the both cases uh pull and push at least for me, I, I, I I, think it makes sense that the same like scenarios or use cases are are laid out in both cases and I.

C

Think the majority of them are the same, but this one is not there like. This example is not there um and I think there is also a non-example there's no example as well for uh because, in the top of the page wheel, there's a a paragraph saying that uh you can uh add the you can add the context on the message as a parent on the process.

C

Pin uh if you want it's not not expecting this convention, because it's not always possible, but if you want again so if I think it's because it's saying that I think we should also have- or we could have another diagram here where there is a uh you can also add the parents, parent relationship to the publishing one and Publishing M2 here from from this uh and also say something the same like better, not covered by the conventions. But you can do it uh if you want um yeah but I.

C

Think apart from this I, don't see any major comments anymore. It's just needs now at this point uh in grammar.

B

C

That makes sense yeah, uh so I will discuss this with with him once he's back, but I think um after this I think the old type is I. Think it's pretty good to go, at least at least for me for me, yeah.

B

That that's yeah, as you may have seen products once. Even if the group comes to an agreement, there's chances there's a big chance that there will be a lot of feedback. So more.

C

Clarification will.

B

A

But it's good that.

B

At least the group you know is happy with the initial version.

A

B

I also have some small questions for him and I will go. You know um turkey gun actually to read on the feedback that he got from all of us. It's mostly minor storm. um That's it. You know so yeah, but I agree that the IIT in good State I was a little bit uh curious about the process supports, but yeah I think that uh the process process.

C

You you cut a little bit: okay,.

B

C

B

Guess it makes sense for now to not include them in the spec, um given the complexity and yeah.

C

I think the the problem with the process is that uh I think we will never be able to fully put them in the, but I mean we can do as he did already here like you. Can we can like somewhat casually say?

C

Okay, you can do this, but it's not covered and show how it would look like if you do it, uh but I think we can't like put an aspect in a way because um it depending it depends on a lot of things, but it depends, for example, on the library that is, uh for example, like in the messaging library that you use uh uh because, for example, some libraries you say like, for example, in the pool based case um uh you just for example, you're using I, don't know the Azure, I, know SDK or or event about SDK, whatever some some library right and then what you do as a developer, say: okay, get messages um of this queue or this topic or something and then it just gives you and then from that point on is, is already over from the library right.

C

So um uh so the library can't can't create the process pan manually uh automatically for you, uh because it's a way out of the libraries code, but someone libraries they they work in in in a way where there is like callbacks.

C

um So you say like uh get messages and here's the Callback to process them uh and that in that case, then the library could do some like injection and and create a span for you for for each of of the processing and so on.

C

um So it highly depends that this this was uh very interesting when uh meeting what we had when Amir uh was joining and he he wrote.

C

I think he's the author of I guess a couple of of messaging sdks in JavaScript he wrote several for several systems uh and he was showing like demoing uh the difference of each and and how this is problematic, because it's some, you can't really uh can't really do anything, um and then you would have to require the user to do it, for example, to create the process fence um so like, for example, if we want to be enforced that there there will always be process pens, then we would need to enforce that uh developers like the writing.

C

The applications um will always create this and I not really sure if this is even possible to do um yeah.

B

Yeah yeah, yeah, yeah I think that, based on some libraries I think that with so for some of those ones we can. We could create this out of the box for the ones that you said it's not possible, so the developer would have to do something I used to I'm interested in this. What I think that could be like version two or three, where we say you know like if you're, if like there, are potential ways to do this, uh and you know, like general advice, you know, and how you can do this.

B

You know, or even right, like Library by Library case uh for now. I can imagine that this is enough, of course, for now, um and we can iterate later and some of these things that that you know they said that this is not covered by the spec, but this is here probably we can include that in a parallel uh docked,.

C

B

Like you know, General guidelines also yeah, like that. It's like yeah, you know like participants, you can add the but Advanced scenarios, and this is more just how it looks and that's it. You know for that.

C

Yeah yeah I think that will make sense. Yeah yeah have at least something that uh if they want to do it, if uh there is a guide, a guide on how to do it and which attributes to record and so on, I think that makes sense yeah, because otherwise, then it becomes uh like everybody does their their own.

B

Way, yeah exactly yeah exactly lane, and then you have to basically merge those ones and it will be a breaking thing for people who are trying to do something yeah. So even if we cannot unify all of those things for now at least you have some of the details set up, so we don't cause a bigger problem later on.

A

C

um Okay, I think I think I think there is um so to answer here. um I, don't think. There's any blockers, um but I will write. I will write here what we went through today and what I think is still missing and then uh Johannes can can pick it up. Yeah.

B

uh Yeah one thing that probably we can have a comment in the agenda is a comment from my side that uh the TC was talking: how to make working groups work better and faster. You know, um and basically, like hotels, require uh at least four apples and uh usually, of course, other people are not involved with these, so that can become tricky so because of that, we want to have two approvals, like one DC member per working group, a lot of people who are actually approvers. You know at these two so uh yeah um to.

C

Approvers uh per working group, you mean yeah.

B

Correct yeah, and in that case I guess Johannes and lumila would come to mind, but if you're anybody else is interested, we can probably make you know approvers for this group. You know, as I said before this is so the step can move on, but also the PRS that we later send. You know to make to incorporate this into the actual. Spec are not you know, sitting there forever yeah.

B

So that's something to keep in mind, but once HTTP smart commissions are stable, we will have to Define deadlines. Who are the approvers if we don't have them already um and so on? You know, but yeah, so something to keep in mind and then so. uh In the meantime, it's a good thing that members of the working group give an approval, even if they are not actual approvers, because that shows that there is agreement in the group. You know.

A

Yeah, that's yeah, so we should so. We should.

B

Approve Delta right away, we're happy with it! Yes, independently of the value of the validity of the approval.

C

Yeah now I take that that external census I think that's what we have been doing uh already like uh when, when we open PRS and uh yeah, all of us uh approve I, think all of us should be in this. um uh It was something that I think that did really in the beginning. When we started uh there, is this inst working group group or not sure how that causing GitHub um yeah, but there's this group uh in GitHub that we are a part of it all part of it.

C

uh This is you can spell like this.

B

Oh yeah, the instrumentation group correct this.

C

Yeah, so we are a part: I mean Johannes, um uh Dwayne I think we are all part of this because we all got added by Dead uh at some point: um yeah no, but uh we already I think we already do this in uh for your TAP I think we just need yeah to now finalize all these small things and we all approved and then I guess we can go forward. Yeah.

B

Probably we'll create any group, probably just to keep things more clear. Oh.

C

Yeah yeah because like because now we are tagged to all instrumentation right and.

B

Yeah yeah I will be like you know, um messaging, that's.

C

A good question or.

B

Something like.

C

That yeah yeah follow up on this and we'll put it on.

B

uh Well, I think we already know uh but yeah you can put a reminder like uh just keep in mind. Like you know, start thinking about uh approvals and the organization already issue will happen.

B

um Http stability is a blocker once that happens we will have to, and the other thing is is that we will well I, don't know, I, think it's a good idea. You know that we had meetings on Mondays, but nobody was showing up, but once the HTTP part is done, I could expect that people can free up Cycles, so we can meet more often two times a week is, is a good one, I think.

B

So in that case we would also have two meetings. You know and that's what the HTTP group is doing at this moment. I.

C

B

They're meeting 12 or even three times a week.

C

I think three times: yes,.

B

Right, yeah exactly so I think that we will have to do that as well. So it's just people get ready.

C

B

Once this happened that that's the next step.

C

Right, yes, actually hard to find uh yeah the time the overlaps with everybody but I guess if there is at least one meeting that happens uh outside the time zone, then it should be fine. I think the.

B

Other thing is that we get like, even if not everybody's there, but at least there's initial discussion. You know, like I, don't know like one one that is more, it's slightly more, uh not European friendly another. One of these is slightly more Europe friendly, for example, and then one like these, where all of us are expected to come, you know that that would work, probably.

C

Yeah all right but I, think um yeah.

B

C

Think if we don't have anything I think we can yeah yeah.

B

Let's see how that goes, so, let's see I'm honestly I'm a little bit uh not concerned about cautious, because I know that things tend to get slower during summer. uh I wonder whether we will be able to finish all of these before you know. Let's see, let's see how things are are going there. You.

C

Know like before the summer I mean.

B

Yeah correct because I remember people taking long holidays, uh even myself, I'm tempted to do that. So so that's that's the thing. You know something to keep in mind, also how this may affect you know things um yeah, let's see: okay, perfect yeah! Thank you. Thank you. So much then yeah. So uh so.

A

C