Open Telemetry Uncategorized, 3 Apr 2023

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2023-04-03 meeting

Description

Open Telemetry Meeting 1's Personal Meeting Room

A

B

Hey how's it going man you're, not your like Studio thing,.

C

No um I went to srecon and all I got was a stupid, coveted 19 infection. So we.

D

Were child careless, I.

C

Don't know everyone on your side to make it out. Okay,.

E

um I uh Kyle go over there. He did not say anything. Liz was there too I've not heard back.

C

D

C

Saw Liz and charity were there. We talked.

B

I haven't heard anything from anybody about it. Yeah.

C

um I only know me or other person so.

D

E

C

Juliano, how.

F

Are you long time no see this uh two weeks meeting is like yeah, I I miss you guys. Oh.

C

I miss you too Juliano.

E

We have no PRS look at that. I know. I was just looking at that.

B

I finally got an open shift instance running locally in my machine. Did.

C

D

C

Tried and it doesn't work the.

B

Default configuration for the VM is nine gigabytes yeah. Our demo will not install in the default configuration because openshift consumes like eight of those nine gigabytes, whatever it was doing. I think I had bought mine up to I think 14 000 megabytes. So what is this.

C

A

Tried the camera trial on.

C

That try other dedicated trial and it won't install the I. Couldn't it was something about permissions um or I, didn't have service account permissions or whatever yeah.

B

There's two definitely there's two things that are failing: I got that far um I was going to kind of work on it yesterday, but I was just kind of just so it yeah.

B

um I was just I, just didn't care right, like I've, been at this for a while and I just didn't care and I was like okay, we I now have open shift. Working. I will look at this later, when I could think about it. A little harder but I.

A

B

Is pretty trivial? It's.

D

B

Needs to go through and figure out which um Services require run as user making sure a few more things, I heard, we might have an issue with the fraud detection service because of how we changed um from a standard image base to the Gradle one, and there might be some permissions issues on the agent. But let's try and work all those through still. Okay,.

C

Well, my my gut was to be like you know what we don't support. It.

B

There and let's make the call on Friday yeah, give me a couple more days at it and we'll try to make a call on Friday for that, um but uh great kind of go: hey Carter.

G

Hey guys how's it going.

E

G

Gone what's happening so.

B

I don't know if we want to talk about anything else, but releasing a one. Four ahead.

D

Of time I mean there's the. What did you do.

F

D

There's one thing.

F

That I want to to talk with yeah go ahead.

C

um So I was looking this morning and there's the the front end, PR or the front end. Issues still seem like they need some.

D

G

Well, Martin's been kind of knocking them down, but he's going to be out this week, so we shouldn't expect any other front. End bug fixes this week. Unless someone else randomly comes in and does them.

F

uh I think he actually he and Pierre they they talk. They talk about the the most important ones so, like uh the the last ones, are just like image resizing. Well, there is one that is actually a book that is the the the math. So whenever I think just the I thought we fixed the decimals.

G

That one got fixed if it was like the comment book or whatever it was.

C

I'm, looking at 327 the I see, five of these are still in fact. Product images are cropped when orders completed, users should get the order, ID and email them checkout view whenever more than one product is bought. The checkout page is shipping data listed multiple times also in the checkout page. You'll need to see more of the different colorants clickable, but the mouse ever behaves like a text selector and not like expected for a link and clicking on the see more. The product image gets distorted.

G

I don't know if we should expect him to do any of the image resizing stuff. It would probably purely be like what.

D

Fixes yeah: what is the.

C

What product images are brought.

B

I think all of them get cropped. When you click on I, don't know, we should double check all that, though yeah.

F

When you, when you add the item to the cart and then whenever you go to the cart, that you'll have like a weird resizing and that just messed up the whole image and each me images in one size, so each image will uh render in a different way but I I. This is just details like.

B

Yeah I think the important ones were the math. um You know shipping listening the shipping data might be something it'd be great. If something could tackle that one and the rest of stuff, like you know, getting a a eye pointer instead of a mouse pointer I'm, not going to cry about that or or a hand whatever it is. You know I'm not going to cry about that at all and images getting distorted or crop like I'm.

E

Like okay, it'd be great.

B

But it's it's not necessary, like images get cropped on Etsy, let's be honest right, so um yeah I I think just uh I still want them fixed, but I'm, not going to say Hey. You know we have to do these to release one for.

C

um Yeah so I think we can probably take the things that.

C

Aren't done and then break them out into individual issues and close out the Omnibus one. We.

B

Could probably do that I.

C

Think the images want.

B

You to be combined into the same one.

C

Yeah sure I mean I'm, not super pressure about it. Just okay, I just want to make sure we're having some level of good hygiene about.

B

Projects and drop the V1, the v14 label,.

C

Well, no well yeah, like anything that we split out doesn't get one for it. Just gets yeah it gets a little bit and then we'll close out 327..

C

um Let's start that bottom actually, uh 233 license linting I'm honestly I feel like this can be a won't fix because I'm not sure that there's any.

C

Like the demo like, is there an actual copyright um argument here like do? We actually need to give a about the license text, because this is not like it would be good to do or we should do it, but it's not like, like this. Isn't redistributable software people aren't integrating this into something else.

E

C

B

Mean I certainly hope would they would they copy paste our code verbatim and then use it somewhere.

C

I mean they could but like.

C

The putting the license but they're not using they could copy paste codes, if it's sure but I, don't think they would be using.

D

Okay, now you've you've.

B

It's literally a Showcase of how to.

D

um Okay, well, let's, let's keep the uh code owners.

B

We kind of need the SDK first off it's funny, because we have the Java team took over the code owner the um the ad service, but they're, not code owner and fraud detection service, which is also Java.

D

F

I never saw them actually commenting and approving any of the that has changed.

C

So that is my point like: are we doing work that is just going to spam people's GitHub notifications and not actually result in changes, because I don't feel like the people that we've added to code owners are really like taking ownership over it?.

B

We should ask them what they feel about this: um let's, let's punt this one out from b14. Clearly we're not going to get it done, nine other sdks to to get onboarded, um but.

G

I think that's a.

E

B

Question to ask to ask if he thinks this is valuable.

G

Yeah I guess I.

A

G

Well, ask: what do you think the Java team, in particular I mean Trask and Jack, are both either on the TC or the GC and they're also involved in a lot of other projects around the hotel, so I think it's a little fair or not at least unexpected- that they're not um no.

C

I'm not I'm, not playing I'm, not no I.

G

D

G

Not saying I'm not saying we don't blame them, but it's also like I, think from our expectations. We should, you know not be surprised and probably set low expectations for maintainers and stuff, like that.

B

Maybe maybe we should um ask who would be somebody who um I I? We should revisit the need for this and and.

C

Yeah I think we actually will.

B

Have support for it.

C

In my mind, the actual goal here or if I had to like put an actual goal behind this, it would be that main, ideally maintainers would be responsible for coming in and updating dependencies in the demo when they release new versions.

C

Right like that's, because we're out of date on a lot of things, I opened an issue like, for example, I opened an issue in the collector of that for this, with the collector, where there's actually like breaking changes with the collector and spam meters and stuff like that, and we, you know, there's just a certain level of IDK.

C

This gets updated every month and I don't know for really. If we've been good about taking those updates um and in a perfect world, the other maintainers are like. Oh yep, you know we cut a release of whatever so we're gonna go put that into the demo now I feel like. This is something that we should come back to, but not super urgent um or it needs more Community discussion.

C

So next one is 606. Hotel operator, injecting instrumentation. This is a big refactor, p issue.

C

B

This is wishful thinking. It's great I.

C

B

It's social thinking.

D

I think it's just I, don't know which.

B

One thank you for one for sorry, yeah yeah. It's we do this. We need a plan to do this right and it requires us to change services somehow to make it work.

C

Ed yeah um crashes back off open shift is that in your purview right now,.

B

Yeah, it's my purview. Let's keep it there. I'm gonna get done.

C

um Error in HTTP server, duration,.

E

F

I, try to update that, but yeah I think you have more inside PR.

B

Ank, um it's I chase the wrong thing, I'm, sorry, and then we let it sit and I. Let it sit for like four weeks and then I up, baseball um I think it's anotherboggan.net. By the way they had the bug twice in their code, so I fixed that one for them, but the one that was actually hitting us was in JavaScript.

B

D

F

Well, I think I.

B

Just got to prove that just got approved last night or this morning or something like that. Will he release I, don't know, maybe.

C

D

C

I mean it's good that we're fine I mean the thing is like this is good that we're finding this right like this is literally the point like.

E

Yeah yeah yeah yeah.

A

B

Can't be mad so that PR and a JS repo was approved this morning, I I hit my GitHub notifications at least cool.

D

um All right so oh 200.

F

Do they have a release process or maybe I can just ping Danielle raila and say hey? Can you maybe push a new release there.

B

If they can we'll take it right for sure yeah.

C

I would be surprised if they like this, isn't a security thing, so I'd be surprised if they do a point release just for this, but.

E

D

uh When was the last Js.

C

B

Yeah, but maybe we get sucked up in the planned release, for you know, look kubecon's coming up. Everyone wants to release before kubecon right.

B

I had to fix core for this. It was four days ago.

C

Yeah uh four days ago, uh I mean I, don't know about it. Oh Daniel, he might.

B

D

C

It's worth it's worth, paying him yeah.

C

Oh I, see 37 times open.

B

Yeah, it was approved, not merged, so.

C

Identity, uh uh yes, well, this is a Prometheus problem and not a.

C

H

D

Annoying thing is actually.

C

Like a Prometheus issue, if you, if you put this in, if you did otlp export or whatever it wouldn't be a problem.

E

B

It I think Prometheus is, should not be ruling against us, but I.

C

Mean Prometheus should be doing a lot of things.

B

Like our Prometheus exporters to not be breaking on this I.

C

Mean I: don't think that this.

B

Exporter itself,.

C

Yeah but that's because of Prometheus like I, think Prometheus is just that way: I, don't or maybe actually I, don't know.

C

I mean I think I researched this right. What is the, uh what did I say.

B

You want you to comment on that one and let them know this is.

G

B

Open television issues permit this, but we just have to work in a bigger ecosystem. So thank you for something that's fixed. Can we get a release? Early I, don't know be great.

C

um So the Prometheus exporter is not merging them correctly. Yeah I, wonder if that's actually a.

C

I wonder if that is actually a Promethean.

C

That might be a good thing to go: ask on them collector.

D

C

Memory leaks, long-term testing.

G

You might want to split these into multiple, since it's three different Services, very aesthetically, two at this point, it.

B

C

I mean: is this a.

B

The low generator one I did a little research on it and it seems like they almost that does this, because they just collect stats for every request and I. Think that's what it is. It's just the internal stats that they collect. Yeah.

C

This is just it's Lucas, just doing it right, Lo and Lucas doesn't have a GCE process right, I'm.

G

C

Some way you could.

G

Tell it to well python has a default GC process. You can just use I! Think it's like some GC library. That was.

C

Kind of bouncing.

G

C

Is there a way in Locust to like reset the stats, every 24 cards or whatever? Maybe.

B

Just I don't know.

B

Is there a way for locusts to like run a test for a specific time, stop itself clear its stats and start a.

C

New one I mean I, don't know I.

C

E

Running configuration options.

C

I mean they have a. You, can run it for stole long, but.

C

Distributed mode.

C

That's stable, printing stats.

C

Resets that well.

B

Let's come back we're going to break this off to two more vrs.

C

Yeah bring that out and then drop the one for I think um because the worst thing that happens in both of these right is it hits the memory limit and restarts, but yeah Locust, at least that doesn't really matter, there's a slight dip in them, maybe maybe a slight dip, whatever all the connections get reset. But it's not like the end of the world email service is.

C

That's the one because that was an Hotel one right.

B

The email service is I, believe grpc, specific I, struggled with my hand on it. I tried upgrading everything to the latest I'm, not quite clear, on how Ruby does gems and package dependencies and stuff like that. It seemed like it might still be using an older version of protocol I think it was fixing like something.21 and one dependencies is using.19.

B

So I don't know if it's using that for everybody or what but I think the bugs in actually in grpc protobuf something weird there um grpc like I I, don't know: can we try right wiring this up as HTTP to? If you still have it.

E

Because it will kill us to do that.

C

There is a that's helpful.

B

F

I did remember.

B

um I tried one time, just removing all Hotel instrumentation from the email service and letting it run through. It still had a memory leak, but it was much much like its curve was way longer term. That's why I'm thinking it's it's a grpc thing because now you're not exporting through the hotel thing anymore and that and so it wasn't creating more.

C

Right there could be just an email leak in the yeah um I think we should anyway, I think that just gets punted.

B

Yeah I think it'll get punted.

D

B

Would love to try to if we could convert that to http http across the board on both its comps and it's otlp export to see if we still have a leak there I don't know what that would take. If anybody knows who Ruby could pull that off um check out someone making a call to it, I'm sure we could get checkout to use HTTP instead of jrpc to make that call I'm sure it would be overly difficult to do.

C

That um all right, so, let's punt that break it out into two things and we'll go from there.

C

um The helm updates. That obviously has the.

B

C

B

Then update is um just waiting on an open shift. That's it yeah part four.

C

I see you commented on this.

B

Here um it just started happening, but I was able to validate it. It just started happening. I was like what the it used to work. It.

C

Did they pull like a new version of the demo container I'm, not that demo container sorry, The, Collector, container I.

B

I yeah one of not too far ago, we updated all dependencies on the helm, chart and I. Think it's since then that it started to happen because we pulled a newer version of The Collector The Collector itself stopped binding on port00 right, right and I. Think that has something to do with port forward two color port forward. Now breaks um stopped working.

C

Yeah, this does feel like something that we have to fix: yeah yeah, but.

E

B

C

Isn't awful the network policy one Tyler reviewed him in The Groovy? Yes, I'll.

B

Poke Tyler, if it needs to prove it, but he.

D

C

He's looking for well, there's.

D

There's comments on it: yeah.

B

um The other one is kind of trivial to do. It's just um specifying.

B

um I'm sure was it.

B

uh There's like they're gonna, do some notes in the uh like read me notes on how to get around it and I. Think it's pretty trivial for us to pull off.

B

um We only need to override the otlp exporter. We just need to specify 127.001 for it. It sounds like.

H

B

I think we really need to do it for the HTTP one as well. We don't have to do it for grpc.

C

C

Oh I see you know, you updated the.

C

Collector version.

C

The connectors should have been moved, connectors, I I just did the notes on this. They should be in beta now.

F

Yeah but then you need to to enable um ah well.

C

No, they it doesn't need it shouldn't. It should be Bomb by default. Now.

F

um Well, maybe the docs.

C

Itself, yeah, but I just did I mean when I. Let me double check my post.

F

But the the precation yeah I saw your post actually, but the deprecation actually is not merged yet so it didn't break anything I.

B

Think it was 47.. Yes,.

C

Okay, no you're right, it did hasn't deprecated it or it hasn't been removed. It's been deprecated, so it won't break until it's removed.

B

What I think it is is I took in it's: when did the home chart change from how the bindings work.

C

uh No we're talking I'm, sorry we're talking about spam metrics now here, oh.

B

Yeah, that's the difference. Yeah.

C

Yeah, no, no you're, good, sorry, I I, just I jump, conversational topics.

B

Yeah, that's that is not one. Four I think it's a post-cubecon thing: okay,.

D

B

Understand, that's fine.

C

Then just I yeah, okay, we're good we're good we're.

B

C

But now Docker.

B

C

Should be aligned now, everything should be using 074 for The Collector. So.

C

um I, don't think it's a pro I, don't think this port binding thing is. Is this anything an issue in Docker as well.

B

You don't no I I, don't.

C

A

Shouldn't be so because yeah.

C

You're not doing the port.

B

Forward so, um and it's an issue in hell, yeah, shooting, kubernetes I- think the fix is just to specify that with 27001 43 18. instead and um oh, we only have to do it for the HTTP side of the protocol. It's the only one that we actually do port forwarding on.

B

um um If somebody feels like Port 40 beyond that, that's outside what the demo does and they can figure out themselves.

B

uh I am also the notion. I think this whole entire. You know if Network policy fixes this I'd rather disable that.

C

Yes, no I think that's fine. It.

B

Feels like a much cleaner fix is just instead of having this weird section. The config, which overrides just one side of.

C

You yeah yeah yeah, no, it's cool so, but it's fine! Actually, we've got a week before we need to like be ready to bang. On this thing, we don't have to release this next Monday. We can just say like next. Monday is kind of the okay I'd.

B

Love for us to cut one four beta images, even if we.

C

Can yeah yeah sure.

E

B

C

A

Okay, so when, whenever.

C

We get so I've gone through. Let's, let's look back at the issues list.

C

I um go to just a 1-4.

B

I put docs in all of the ones, I think we're going to do um um need to fix this still so I filled the docs in there, but yeah there's a couple of them got to drop the labels on.

D

D

um I guess one.

C

um I mean the link the license linting shouldn't, be that hard should it like. Is it? Is there a tool that can do this.

B

I would have to think it would exists and there's a way to wire it up.

C

And get it I'm sure there well I'm sure.

B

C

To exist, I mean it's like that. Let's keep that. Let me I'll I will find some time later this week to look into that.

D

C

Let me assign this to myself.

C

Drop the on the bus I'll give Pierre I'll put you on the.

F

C

B

We're gonna remove 606 as well. Did you.

E

Already yeah, you already did that yeah.

C

B

771 you got that one okay I'm just going over the notes here, because I took notes, as we were talking about all these yeah um okay, yeah 737, we're gonna hope that thing gets merged in time. I think 737 is probably the only one that is at risk because we're depending on something else, yeah and um if somebody wants to do all the helm updates that'd be great.

B

Updates, like like just go to PRS and search for a Helm update required in v14 and you'll, see everything that needs to get done.

E

A

My intent was to.

B

Do yeah I'd love to because um yeah I can.

A

Tell someone else publish.

B

Images for now and I could give them like one four labels and we'll just call them like under my my Docker hub just um for testing it sucks, because after like the fifth time, you hit rate limits on docker because we have that many services.

B

So it's it's yeah. It's annoying.

H

Yeah I can chip in and I do have I I do have it on my. You know on my mind to make that small change for openshift as well and I've. Gotten I went to test that in.

B

Yeah I finally got CCR stuff running you weren't on earlier Josh right, but I got it done. I have open shift now running locally I've seen the demo fail. I seen the two services that came up with failures on security policies.

B

um You need to give If You're Gonna Do It um locally. You need to give the thing like 14, 000, megabytes.

C

um That's that's red hat quality right there.

B

Look I started off and I just didn't like get pods all and I was just like a whole lot of there's always a lot of stuff in here everything everything.

C

G

Enterprise agreed.

E

B

Yeah, but you know we'll go to work because I think um it allows us to get into a lot of organizations that are otherwise built in other ways to do stuff. And it allows us to enable them with tooling.

C

Yeah yeah: no, we should we should figure it out. Oh, yes, all right. We have 10 minutes what I.

B

Probably be great if we could cut a um yeah at least a Beta release, so we get ghdr images published which allows us to get around rate limits on Docker, hub, yeah uh and yeah. So.

C

Yeah everyone drive towards code complete by Friday, so we'll can cut a Beta release.

D

B

I think 777 will be the only one not done by Friday and 7 37.. Those are the two Airlines for Boeing. If.

G

B

C

C

But yeah, so what else we got going on I.

G

Think uh before we go, Collins joined the Sig for the first time, so Colin.

D

G

Have any specific topics of Interest questions? We love? We love to talk yeah.

I

Yeah, hey guys, uh how's it going. um You know, uh yeah, it's been on my calendar now, for for a few weeks, I I pulled down and started. uh You know taking a look at at a lot of the um the work that you guys have been doing.

I

Kudos hats off uh for for the the work, I think it's it's going the right direction, um we're actually having a conversation in the organization right now um uh about um you know, funding um a new engineering team for uh for for demos specifically um for us- and you know we are at this point considering where do we put that that development work going forward? I? Think um right now the The Proposal is is to do this in in the context of of hotel.

I

I will say that you know in terms of I think where we're going um this this notion of unification of desec Ops. We have you know Cloud scene, Cloud, soar and security, and so one big piece of this is is around logging right. So um if we start to come in and do some contribution, I think some of the things that we're going to want to do is, you know, see more problem patterns and and do loggings and those sorts of things and and have really good log streams being admitted yeah those those sorts.

A

I

But but we're not there yet I, you know I just really wanted to kind of jump in and kind of get a sense for where you guys are at um and and what y'all are working on. So I can kind of try to see how that interfaces on some of the stuff that we're talking about doing.

C

Yeah quick question which org sorry Sumo, Sumo logic, Sumo, okay, great great yeah, um so logging is actually the next like decently sized bite. I would say um especially now that hotel, that, like the logs Bridge API and everything is firming up. um So definitely you know I I. Think after kubecon that's going to be end up being the focus um so yeah we would certainly love. There is a lot of I mean it emits a lot of logging right now right, like there's a lot of logs that come out of the.

C

Out of all the different services- and you.

I

C

Would be pretty straightforward to at least you know, kind of a V1 um or or an MVP would simply be having the collector like swapping out the you know, having the collector start to collect those logs, yeah um standard out standard error and putting them into something.

C

The kind of that's.

B

Actually, a good discussion there Austin should we be relying on collector to scrape or should we be looking at the SDK log.

C

I think we need to really I mean I, think collect I mean The. Collector should be scraping logs because that's how it's going to work in any place that does this, like, certainly for app logs as the bridge API matures and like in places that there is a bridge API like I, think that I basically feel like there's kind of two or three distinct patterns here. Right, one is the sort of I, don't say trivial, but sort of the standard way that people do log collection which is I, have an agent or whatever that's scraping.

C

You know, stand it out or files um and that's a good starting point for us. I. Think then there's um push logging is, you know through the bridge API, which is a also a value one, but I think that's going to end up being something that we're going to need kind of specific services for like that, would look more like a.

C

Like an actual for deployed use case for it, like uh I, don't know like a client, you know like clients, you would probably push logs rather than scrape logs right because you'd have to, um and then the third would be just sort of like novel at limitations of the hotel logging API.

C

um So maybe the collector itself sending the collector's logs somewhere.

D

C

I don't know if the collector supports otlp log output, yet.

D

C

Yeah collector.

B

Does support it.

C

Does yeah it? Does it does okay, so great, so we have so we're. We have like two out of three right like and the third, the the the four deployed logging thing would probably be slapping it somewhere in one of the Java Services. um So after coupon or even during kubecon I think we can talk about.

C

um We can have some conversations at kubecon about like putting up a design dock or starting a discussion to get the design out. um I will say the.

D

C

Thing, that's kind of been an open question is what do we actually ship in terms of a logging back end with the collector, because there's no cncf project for it.

F

Yeah, this is a important Point, actually.

H

C

E

Don't want you know entertaining.

C

All entertaining any and all opinions right now, because otherwise it's going to be open, search, right, right and I. Honestly, don't care like or I shouldn't, say: I, don't care, but I think open search is like the safest thought. Is the most open source option.

B

And the most widely used, probably for.

C

A lot uh I, don't even mean I, would say, I think it's probably still the most widely used option, but OS is okay. They're licensed friendly is like the most licensed friendly and it.

B

Can important friendly version yeah, that's right.

E

B

I

Some of the other things that that um that I've been looking at and thinking about really I, I'll I'll put some some uh some feature requests in, but I think one of the other things is is making you know, timings more realistic, like um everything in the micro seconds in terms of processing as like. All these transactions are really really really fast.

I

As far as what's going through the mocks and low drivers and everything all the calls to the back ends if you're gonna, if we're wanting to represent a demo environment, that's going to have like performance characteristics, we should have longer response times, I think in a.

B

Lot of these, my team talked about exactly that um at our team meeting last week. Yeah, um we keep on saying that we're going to dedicate time to actually plugging in real scenarios into a demo and and part of that starts off with putting some um random windowed wait times like this. This call should be between 20 and 50, milliseconds, yes, and and just randomly picks a time between those those spots. So now you have a a trace waterfall that looks far more realistic. Instead of a bunch of things, microseconds.

C

So I agree, but I also don't want us to overload the thank you. We should do this, but we also should do this in a way that is not.

C

I don't want to I I feel like we have a responsibility to the like realistic part of like making it a real demo by not overloading the code with a bunch of like fake, basically, especially in our Brave New World of llms and code, you know and AI code generation and adding things that are like. If we you know.

C

However, we do this, we need to do what we need to do the meta programming in a way that is like very distinct like separates the logic, the business logic and the instrumentation logic from sort of the like. Okay, now, let's add in some timings and crap right, yeah.

I

B

I

Having a harness in place to do that right so that you.

B

I

Can you can you can decide what patterns you want to have and, and you can yes yeah and provide some kind of a mock harness control.

B

When we did this with microservices at honeycomb, it's still our demo because it has those timings in there all the random wait time. Things is a separate function and for every service.

E

H

Don't actually see.

B

Weight inside the function, we make a call to and I think the call is actually mock random weight or something like that right again,.

C

We just need to make sure that, like when we're adding in stuff like this, that we're doing it in a very composable way, um so that we're not um conflating right, yeah or we're not putting in a bunch of like that is going to just make our lives a nightmare in five years when the AIS take over um but yeah.

C

If you wanna I, think probably the next best step would be, if they're, a specific FRS and there's specific things that you all want to see that you're willing to dedicate a sharing time to go ahead and file. Those. And let's get the conversation started um and then, like early, you know, late April, early May,.

I

We're going to be second quarter for us, which is you know, um yeah really the end. You know right. It's really going to be end of April May before we start to spin anything up, I think.

D

Yeah yeah totally.

C

um Yeah so I think end of April early May well uh after cubecon. Certainly yeah we'll have a we'll go through we'll do our road mapping things I I would like for us to plan out like where do we want to be by um scubecon in a yeah as a you know, as a demo, um so we'll, let's Orient ourselves around, like Hey we're gonna figure this out, we're going to figure out the next six months or whatever um and after Cube kind of you, that's good folks.

I

Well, great nice to meet all of you.

B

I

For joining us, so so yeah we've got we're going to be there. um We I'm I'm in the US and and we've got most of our media folks are going to be there um I'm still, apparently we were uh relied on getting tickets and tickets sold out so there's a waiting list, so I'm I'm still waiting to see if I'm gonna get uh the lottery. There's there's two or three tickets left and we're trying to organize and figure out who's going to be over there. So yeah.

D

I

Hope to be over there.

C

Well, if you, if you get over there, look us up if you, uh if you're, not and there's someone else from your team or someone else, that we should talk to just yeah my way, I know I'll, be there.

A

C

Be there right.

B

I will be there and for those of you who are attending if you have not signed up for it yet you'll be there in the Tuesday before please sign up for open observability day.

C

um Come join us also. That is also full yeah.

I

C

C

If you will be there uh we'll be doing uh a reception that night uh Tuesday night so uh come find me and.

B

Yeah absolutely.

I

B

Might open that reception, early Austin on our own tabs, and maybe we can just.

A

C

F

Just text you uh the reply from Daniel on the slack. So if you could take a look he's waiting on that to actually do the the Virgin release.

A

B

Oh um awesome, chime in here on this thing. If you want that's exactly.

A

What he's waiting.

B

For he wants to know about the um did you DM me, then what his question is. This seems like a um on epr for the JavaScript, the pr for the JavaScript bug. He was questioning how he's questioning well, it should not be identifying, but this is really a Prometheus exporter bug.

F

I can't actually.

A

F

A

C

F

A scene I was seeing ice cream.

C

3710 and hotel JS yeah.

B

So I'd love to get an answer for that.

A

B

F

D

Thank you awesome.

F

C

Yeah thanks all right, I'll comment on this right now see.

B

You guys um wait. Our next one is a Monday of kubecon. Are we actually gonna have a meeting.

C

B

um I'm gonna be on a plane, so no no.

C

We're not I'm speaking, it rejects in a bunch of our so we'll cancel that one and maybe we'll have one next week, I'll post on the channel, without it, okay.

F

C

F

Right all right uh enjoy, it could become guys, see you bye, bye,.