Layer5 Meshery Development, 30 Mar 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Meshery Development Meeting (March 30th, 2022)

Description

Topics:
- Discussion on Meshery Compatibility Matrix summary.
- Review of Meshery Adapter for Cilium beta readiness.
- Finalization of Helm deployments of Meshery - does mesheryctl use helm dry-run or helm update?
- Revisit Meshery v0.7.0 feature theme champions

Join the community at https://layer5.io/community

Find Meshery on:
GitHub: https://github.com/meshery
Twitter: https://twitter.com/mesheryio
LinkedIn: https://www.linkedin.com/showcase/meshery/
Docker Hub: https://hub.docker.com/u/layer5/

A

Wherever everyone is um yeah glad to be here, I first meeting um I uh yeah, I was introduced by a friend to their five uh had a chance to meet with lee.

B

A

Chance to meet with everyone happy to be here.

C

That's great, thank you for joining in michael.

C

I guess that covers the newcomers today. So let's get started off with the agenda uh aditya. Would you like to share your updates on the compatibility matrix.

D

Yep sure I'll try my screen.

D

Just share the design mock-up, which we, I am very, very discussed actually.

D

Okay, so uh our task is to create a compatibility matrix for the measuring adapters, so that which connection is unstable. Okay, I don't know what it is sorry for the interruption. So um for two days uh I and subedi were talking about how the compatibility matrix would look like. So it's like basically uh collecting overall the test results of the integration test, as you can see here so like for each times tab, the tests are being run and then the results are being posted and then published onto the message.

D

So our objective is to create an overall table which represents the um the overall results of the measuring uh test for the mystery adapters. So this is actually the design mock-up. So I know it's a bit of a bit skeptical, but what we had in mind is like it will be this all colors will be in the form of symbols like for green.

D

It will be a tick mark for or yellow it will be, a tick mark with the yellow, color and for red it will be across, and for this one uh I just referred like some of the measuring adapters are not being tested yet, for example like console and sm, and so I gave this color code like not tested, so this is just a overall markup so uh and also we are considering like this, what are all the stable versions, pressure present and the edge versions, and therefore, what will be the order?

D

Compatibility like taking the probability of this two case and then finally, the result will be appearing here. So since this uh both are good, the world will be good else. It will be false. So are there any cases to be noted or any notes to be noted so that we can do better on this.

B

um Alpha did, did you get? Did you show the existing um compatibility matrix? I I you might have. I just didn't see it.

D

Yeah, just I just saw, but unfortunately the link um is forgotten, so I couldn't open it right now, so I'm just showing it. This is just what we had, though, and we discussed- and this is just the.

B

So, oh yeah, no! This is good yeah! This is good. Actually what I mean is um the other docs page that you had up it's a good like like the mock-up that you have that information.

B

That summary information would go at the top of that docs page. So, if you would, if you go back to the compatibility matrix.

D

You mean this one.

B

Yeah yeah, um and but I I apologize, did you show this when you first started talking about this topic.

D

No, no! No. I just wanted to show this one.

B

Great great okay, good: this: um do you want to introduce people to this, because it'll give a lot of context as to uh the this unit of work and they'll be able to offer a little more insightful feedback? I think so. Do you just want to introduce people to to these integration test results.

D

Okay, so uh for guys uh who are new to this compatibility matrix, like so ashley mentioned, each of the measuring adapters will be made to run in a separate environment, and then it will be tested like whether this test results are being passed or not. So there will be some integration test being run by the gita factions, which is present in each of the some of the battery adapters repo and to view the past results. If you click this uh click this, then you would get the whole history of the.

D

What are all the tests being done and been at during which, at which duration, the test of being run? And if this passed or not sorry for the so yeah? This is for the measuring istio and, as you can see here, for the measuring component version like there are two tabs reconsidering two cases like edge and stable, so stable means where it will be denoted in the form of the semantic versioning. For in this case version uh 0.5128 is the stable version, and hence it will mention here if it is being tested on edge case.

D

Like say, if you introduce the change in fpr and it is uh immediately merged, then it will be considered as an edge case, so the edge version will be considered, so you can actually view it, and if there are any tests which are being failed or just inappropriate, then you can just see them over here, like you can get that overall info, like okay now in this case like so some more not able to being run. So in this case, it's like this.

D

You, a agress gateway and same way like that, so yep, that's much of it.

B

Nice, and so what alpha is working toward and along with su betty who isn't on, is yeah. It's just a summarization of this information, it's great to be able to click drill down into the details. There's sort of two levels of details for each of these test results.

B

um The page itself, the title of the page is is called a compatibility matrix and it it it's something of a lie. It's something of a farce because it's, this is really a set of end-to-end integration. Test results, uh okay, so where's the compatibility matrix. Well, it's somewhat. Some of that data is inherent to the results. So is the meshri istio adapter compatible with istio 1.13.2?

B

Well, it's green! So it looks like it and actually it is and that's what those the green check boxes mean: okay, uh good! What about kubernetes? What versions of kubernetes is mesh three compatible with well actually for the istio adapter, it's green against. um You know this version of kubernetes and as like, it's not just kubernetes. It's not just the surface. Meshes it's um well! It's a really complex! It's combinatorial math! It's! uh You know it's mac, os linux, os windows. It's x64: it's arm.

B

It's m1s, it's just a long list of it's well linker d, stable 2.11.1, what about 2.11.0 and so on and so on, and so there's a lot of you know. So the the test framework um and the the notion that management software like mesheri is compatible with the things that it's managing is quite critical to the value that mesh re provides.

B

And so the thought here that alpha is working toward is well a useful summarization of that information. There's good inspiration that you can find from well. Actually, I think the screenshot um that's in this issue is vmware's compatibility, matrix, it's uh probably more than what is needed here, but but yeah. If someone comes with the question of like hey, what version of kubernetes is measuring compatible with, and it's not just messaging, it's mesh reversion such and such.

E

B

So you know you know everybody familiar with pivot tables like uh any anyway. um So with that context, I'm alpha if you go back to your um the mock-up here. So if someone comes just under that use case, they come with a question of was measuring. Is the latest measure compatible with istio 122.

B

And we would probably want to drop an entry in here for kubernetes.

B

um What else like yeah, so we also have sni conformance test results if you really wanted to complicate things, but I I don't think that we we do at this this point, but but you're hitting it about right. You know, at least of the feedback that I can offer, is um what you might so so you've got individual. What you might want to do is given that there's more info to to present. Maybe we would hide the edge versions for the moment, maybe maybe not, but have meshry server as a component.

B

Meshery perf is an upcoming component adapter as well.

B

Yeah one of the things that uh there's I don't know if um sumiya is on right now I don't think he is, but one of the things to give consideration to is that this compatibility matrix that you have mocked up here.

B

Good, like you, know, good, so it says for for a given version of a measuring component here, here's it's.

B

Here's it see yeah, you need a mul. You need kind of a multi-folded, a multi there. You probably need a drop down at least maybe one or two top level drop downs or top level um view switchers. That would change you from the latest measuring release to a release prior and like what does that? Look like uh okay uh for like because if someone says, if someone comes with the stipulation that says, hey, I'm running istio, 1 18., I'm sorry kubernetes 118..

B

What version of mesherie is compatible with that that they they may need to switch.

B

Through the they need to hey boy, did: can you can you go to mesherie.io, slash smi, maybe we'll give a different example of this.

B

Does anyone else have feedback by the way? So so it's it's a little bit of a pain and rump it's um information architecture and how to best summarize and present to users, people would come and say: well, I'm running open shift version such and such and I'm running it on.

B

Well, probably red hat linux right um and they they'd want to know, does does measure compatible with that so yeah. If you were to go there, here's a different example, relevant example of measuring, performing tests against individual service meshes and the different versions of those service meshes. It runs these tests to see if.

B

The service meshes are compatible with smi smi version 0.6, that's what the dot 6 is in the upper left hand corner good what about version 0.7?

B

Okay, so if there was a selector there like a drop down, then, given that this is a service mesh interface report, you could then browse by service mesh interface version who is compatible like that's um yep, so now now that that that I think, is mostly making sense. Let me let me make it more complicated, which is that well like in this example um great so there's different versions of service mesh interface and the four specifications that service mesh interface has each individual service mesh had different versions. Are those versions compatible or not?

B

Now the test cases themselves? So so, why there's a green check mark or not? There are specific, there's a specific set of test cases actually yeah. I think if you click on one of those like, why did it fail? If you can, you is that clickable by the way? No.

D

No, it's not clickable.

B

Okay, but if you go down to the bottom in the legend um there, it should be a link that explains a bit more beh so that that hyperlink at the bottom explains a bit more about the test cases themselves like what tests and what assertions are being verified specifically. So if you scroll toward the bottom there's a list of uh yeah, so these are the specific tests. So so what makes this like one order of magnitude more confusing? Is that the conformance test itself has it its own version.

B

Today there are 35 tests or some certain number tomorrow, there's another there's potentially another set, and so we don't have to figure that out on this call.

B

But I was just helping expand the thinking here, as you consider how to summarize so, okay, so who's saying uh so, there's a couple of folks at least that I'm that come to mind off the top of my head that have either been at vmware or are currently at vmware, and so like husseina and and michael. um This will make probably a lot of sense to you as as well as to the rest.

B

But if you alpha, if you were to open up the github issue for this particular body of work, this screenshot in there should show like just as a random example like vmware's compatibility, matrix.

B

Yeah so just you know, here's.

B

Here's your online pivot table or, and so hey look. We don't have to get there today. um Okay, I think you have a general sense, like the mock-up that you have is is aimed in the right direction it. It is a summary you are presenting a summary.

B

You know, you're on the right path. There are others on the call that might have something to add.

F

B

F

Yeah, I just have a suggestion, like the compatibility metrics, I think adding something about the operating system on which the tests were run and whether they failed the past.

F

Would I don't know I I don't know if it's crucial, but from my point of view I think it's a perfect candidate to consider for inclusion in the in the matrix. Also thank you.

B

Alpha, what do you think uh who's saying it? Do you have a perspective on this.

G

I believe what security is saying is that we should also include os compatibility like windows, mac and linux, based.

D

But generally are we considering the os compatibility for the in this case yeah? It would make it complex right so.

C

C

Yeah os compatibility, we probably shouldn't be putting in the same table. It would.

C

D

C

To represent it right.

D

Yeah, it would be difficult.

D

C

B

C

Any more does anyone else want to say anything about this. Any more thoughts.

C

Okay, I'll take this island says no uh thanks, avithya for sharing this. Let's move on with the other items who's that would you give us a small demo on the psyllium adapter.

H

Oh yes, sure I will share it on the screen.

B

So yeah, so while there brings that up just a little bit of context for everyone, so today, mastery has 10 different service. Mesh 10 different adapters for individual service meshes um pop quiz.

B

Why does why does mesri have an adapter per mesh? Why doesn't it just have a single service, mesh adapter.

B

There's there's it's like extra credit on this pop quiz.

H

Is my screen visible to all yep.

H

Just a minute.

B

So um while, while you all contemplate that question about why they're individual adapters I'll fill in the time here and say that pranav is on to something in the chat, I also say that there are 10 adapters. Currently, one of them needs to be turned down.

B

Speaking of vmware, it was the octarine service mesh that adapter that company that was acquired by vmware some time ago and in that time, uh that mesh has that software has gone away or been reincarnated somewhere else, so that that adapter needs to be um deprecated um and so in its place. So we'll still, while we're adding another adapter for psyllium, um the overall count will probably stay at 10., so we'll have to.

H

Okay, so you can see the uh machete measuring clam adapter serium support for clam adapter has been added to measuring ui as well. You can see all the basic operations of the uh installing the celium uh service mesh and the sample application.

H

uh You can now uh do it with the ui as well and as well with the measuring cli. The earlier issue of life cycle management, page was not. Loading for the celium services is also been fixed.

H

As you can see, I hit the install button and we get a success message.

H

A service mesh successfully installed and the other one of, and also the cm adapter, is now using the latest version of meshgate adapter library and the regression from it of uh we can and the regression from it that clm service mesh was not able to install it's also been fixed and now uh another and another support for component generation for the new version. If available and making it a static, statically available has also been added.

H

We can you, please uh specify what else I can show you or what specific details you want.

B

Yeah yeah, sir, did you were you able to confirm that psyllium is in fact installed on your cluster? Yes,.

H

uh Celium operator and the deployment and the replica assets for this helium adapter nice uh good here is one thing: as you can see, one of the parts for the serum doesn't come up, it always dies. I don't know if this issue is with my local system or or elsewhere.

B

I can uh I can sense, mr talarico, just just just wanting to chime in and uh and maybe share if, if it's common, so the the pot, okay, so you're saying which one okay so all yeah, it might be. The resources on your system potentially.

B

Although is it consistently the psyllium pod that has.

H

Failed, yes, there was issue regarding this.

H

I think this issue.

H

But I was not able to understand why this this is the case.

H

Can you confirm, is this issue with your system as well yeah? Actually, I was yesterday checking it out, and I was also not able to see running for studio, though, if, like.

C

H

I like, if I do, we don't use adapter for installing, and I directly use the hem chart. Then it does get installed but like if for adapter, I think there's some problem. We need to figure it out, though the hand chart works. Fine if we use them separate.

B

Nice so joe talarico of uh the cilium project was he was just on. uh I think he's on another call, but he knows that you guys were demoing, so he wanted to come, see it um and so he's a psyllium.

B

uh Well he's the psyllium representative. So a good he's a great person to ask it may well, you know to, since we don't have a totally packed agenda today. One thing that we might want to just take a quick look at is us. If you were to do a cube ctl um logs on that particular pod, uh there might be a decent indication as to why it's having a hard time coming up um pranav when you do a helm-based install.

B

Outside of mesherie, do you are there any particular configuration values that it requires? No one.

H

Thing is that, like for help installed, celium uses its own ctl, which we don't have right now in the adapter serium cdl. So I was like I wanted to ask this like some. I just it's just for my mind that we, I need also need to add the serium cdl in the install operation so that, if, for example, hence like fails to install, then it can fall back to serium ctl for installing.

H

As as um I think in istio adapter also, we have something like this.

B

Yeah yeah, that's a good point. Yeah.

H

B

Supports that is there a sec? Is there the secondary? uh I'm sorry is: is there the psyllium ctl in there the adapter currently no.

H

In the adapter, we don't have it right now.

D

B

uh So I don't know if you're lucer, I don't know you've been put on the spot. A few of you have been put on the spot on today's call, which is which keeps everyone on their toes. But um I wonder, is: is that psyllium pod inside the default namespace, uh if not knowing the cube system? Namespace?

B

Okay, that if you, if you go ahead and if you do that same command again, but you tack on the cube system, namespace.

I

This command right.

B

Yeah and so yeah, if you do, if you do that same command, but with a dash n uh cube system.

B

And then the name of the container again, the pod.

B

And this is their might.

B

Keep keep going if you would.

B

Logs interesting keep going.

H

You can describe.

B

Yeah abby, I don't know if you were horrible.

C

He was saying if you could do cube ctl describe I mean I, I couldn't see any uh error level.

B

The fatal one at the bottom, it might be the envoy binary psyllium envoy- cannot be executed.

B

G

B

If you take that last fatal message, that's a good.

G

B

One to go research of it.

G

F

B

Joe tolerico, in uh slack.

G

He might be able to.

B

Point you in the right.

G

B

So, given this um uzair this, this is great. I think we're like on the precipice of having a new a new adapter in beta. It's pretty exciting sounds like we've got an enhancement request as well to support a fallback fallback provisioning.

G

B

B

Anybody else have feedback for lucer or pranav.

B

Good, I sense that you all share in my excitement. This is great, very timely, adapter that well, that project psyllium and the and the startup behind it. I surveilled they're quite interest. I was just meeting with them yesterday, they're quite excited as well, so particularly around performance management and how well their mesh performs. So in a weird twist of events, they might be the first one to begin self-reporting their performance test results, which is why we built out the github action.

B

But that's not the only reason why, but it's one of the reasons why and so time will tell first we'll have to figure out why the envoy binary is falling over.

C

Okay uh thanks joseph thanks pranam moving on push. Would you like to speak to your top topic.

I

Okay, hello guys. um I was not aware of that. There was a topic on my name: okay, so, uh okay, I'll talk like I won't be able to share something right now, so uh system start won't, be implementing any kind of try run okay, because that is not needed. Now we came up with some other solution to maintain the helm errors. Okay, right now you, I hope you won't see them for at least quite a few time and uh with a system stop I'm thinking to implement a dryer there.

I

I like to use something that we are implementing uh the force cleanup cluster like uh if, in any case, like messy ctl, does not like helm, fails in case uh to clear all those uh deployments, service, accounts or secrets, and all these details we will. We will try to use the dry run, to check that if it will fail or not. If it is guaranteed to fail, we will implement like force cleanup our if it will like.

I

If it will work, then we will just go our way like use the helm uh method to delete them, so the iron will dry on will be there in system stop, but it won't be in system start.

J

Okay, so I don't know if this is too much to expect it's called. uh Can you tell me how the triad was.

I

Okay, only like on the very basic uh like I, I don't have much detail about how it actually works. Trying helm provides itself a dryer mechanism like it will. uh How should I tell it?

I

I haven't read about it, how exactly ryden works behind the scene in hand package, but it has a feature to like try to dry run any kind of action like maybe install uh uninstall upgrade all these.

I

These methods can be dry run and it can be used to like get get to know that if it will uh guarantee to work or may it may fail like so it kind of give us that kind of outputs. So we can use the like leverage that functionality in system stop in system start. It is not required.

B

Nice, so actually just to clarify and to help everyone kind of be on the same page on this, because it does, it will often affect.

B

um Contributors which, uh which you know a lot of you, are so I'll share really briefly part of. What's I don't know, maybe we'll finally realize the full potential of our choice to centralize on helm, but often what will happen if you're, if you build meshry server and it connects to your kubernetes cluster, one of the first things it tries to do is deploy the mesherie operator. Why? Because you want to get mesh sync up and running as fast as possible, so you can reduce time to value for users.

B

So you can, you can begin discovering existing. You know, discovering your kubernetes infrastructure, your environment and sending that information back to measuring server, okay, great so necessary server will provision the.

B

Mesherie operator and mesh sync, I'm just realizing, like it's probably better to just share some quick visuals I'll share these visuals as I go to ask piyush a question which um which which push so there's actually a set of sequence, diagrams about how it is that the mesh reoperator um is deployed and kind of all that that entails um and uh so to cut to the chase actually, rather than explaining it all um push. Do you so we're? Not you.

B

Mess with ctl is just to confirm, because that is it's not using helm dry run, um but it is using helm, upgrade.

I

Or no okay, so uh how it works I'll, try to explain it a little more! It has no dry run system start has no dry run at all. uh If there are already some deployments running okay.

B

Okay, that's fine! I mean I just yeah, that's fine. I mean I just like. If you're not you're, not using dry run you're, not using the upgrade capability of helm, so you're just manually detect, you know if on failure, you'll go ahead and delete and then reapply and now that they've got the same release name, that's no longer a point of contention that.

I

Yeah, that's no longer a requirement because it will not detect that. Yes, there are some some parts there some deployments there, so it will automatically upgrade its method to install to upgrade like it will switch on the method, so it will switch and then then it works. Fine.

B

Nice good, I only skipped over that just because we normally take a lot of time to um explain and explain- and in this case whether we've explained this one like we're, we're both sort of tightly gripping onto our sanity on this one and uh kudos to you by the way for having the wherewithal to see this through.

B

um I think helm has one two three four letters which I think is pretty appropriate um at this point with our experience uh as a four-letter word, so yeah good. So what piyushi? I think the takeaway from most of us is saying the next time that you're using measuring ctl to either start a measuring deployment.

B

Stop a mesh redeployment, update, a mesh redeployment, restart one.

B

If you see an error, come back he'd like to know about it, because for the most part we consider that those are gone like with respect to this challenge between really the challenge was like a to a split-brained problem in which mesh reserver was being partially responsible for its own deployment of the operator and of mesh sync, while the other brain in this case was meshrie cto, who was being also responsible for the the deployment of mastery operator and mesh sync, and between the two of them.

B

They had discrepancies, and that has been plaguing us for a while. So this really isn't a user facing concern.

B

uh um As much as it, it hasn't been as much as it's been a contributor facing concern. If you're interested in, why that is just you know, ask on the discuss forum and we'll explain it.

I

Yeah, mostly these, uh these kind of issues will be faced by contributors because, like you always deploy and delete multiple times a day and like your cluster, get gets messed up, so users will rarely or sometimes will kind may face it.

B

Anybody have questions for push in this week raj. I know you've been sort of recently um engaged here.

B

um Any other thoughts.

J

Yeah so wait. I just wanted to know what's the problem with uh directly like with what is the problem of directly going and deleting resources like? Why do we need to.

I

Okay, so yeah maybe like, since we face uh deletion problems, sometimes and that's why we created the dash dash force flag and uh it may like it's really much as a frustrating thing. If my ctl says that yeah everything is deleted, but the person can see with cube ctl at so many sources out there and that's why the like try run and the force cleanup is needed in system. Stop.

I

If you do a good job.

J

If we do a good the force cleanup itself, then.

I

Okay, okay, so if we make it like make it mandatory like the force cleanup, it still might mess up because, like we are not doing it like with the uh like helm, whenever install something it it, it does a lot of things additional to the manifest. Whatever the values we are passing, it adds more details to it and if anything is left behind the next time user will install it might get messed up and like he will face any more issues.

I

So it's like, we can't make it like a compulsory one like we will always force clean up force. Cleanup is the fallback will be the fallback step of uh like uninstalling measuring components.

J

B

Yeah just to add to that raj, that's a great question. It's um like unto its own. The answer to your question is like that there isn't there's nothing out like innately wrong with simply just you know, force deleting any residual.

B

Meshry artifacts, you know any residual components left behind or orphaned as part of a deployment, a deployment that's been stopped um unto its own. There isn't an issue, but um as a design principle like pattern, um it's just it's riddled with sustainability issues and edge cases, and- and so in this case we want to use.

B

um Ideally, we want to be using the same mechanism to un-deploy as was used to deploy um now it that said like it can be helpful to have a force command that or fl feature flag that augments the behavior a little more forcefully um do. Did that make it in by the way did we do we cut the that flag.

I

It's in system stop the orange system start.

B

Oh sure, okay, fine yeah- I was that was you guys passed the test. I was just checking to see if nice, so do we okay, great perfect, I can there will be less four-letter words escaping my mouth thanks to pius. I appreciate that any other questions or feedback before we go to the next topic.

B

So I think um subedi was on the call earlier um had to drop off. That's uh that's you, man, you're up.

J

uh Yes, so uh oh.

I

I just wanted everyone.

J

Okay, so this is basically um so the problem. The issue was that uh running mexicl version or error out with this panic, like one of the errors possible, would be if, if your message server is not running- and you run message, ctl version then, instead of displaying an error message that you just panic, so the reason was because we were using the mesh kit logger to log non non-mesquite errors, which is documented to panic.

J

App planet planet applications, so I just replace all errors with all the mesh kit errors with their own error codes, so that not be locked without.

B

Nice good deal um questions for rude raj did you did you say why it is that we, the issue, came up in the first place.

J

Because we were using the mesquite logger and we were trying to log non-measured errors and I did yeah. So I looked through the docs from mexico that.

B

Got you very good, sorry um subeddie? I didn't realize that you were on d. Do you want to do you want to take the next topic? Yeah yeah.

K

uh So I've been working on the card design for the pattern. The for the designs like uh we have a card design on the profiles.

K

Sorry, my network is a bit slow, but yeah. uh I think everyone has seen it. We have a card design when we go into profiles and when you try to manage profiles, we see card designs like these right, so I'm trying to replicate the same functionality to the design, design page as well uh so till now uh yeah. This is this: is the update like I've implemented the card design and the design like there are some networks left right?

K

We have logo missing here and the uh buttons are not functional at this moment like it's, only the deploy button is functional, but yeah, and this is the back view of the card so yeah any any comments and feedback.

C

I feel like it looks pretty good.

C

Just a couple of things: yeah yeah, I think overall, it looks quite nice.

L

Yeah, if you would also show the list of tables by toggling the menu yeah.

K

L

K

Okay, that remains same as this. This remains same.

I

God always concerned about.

B

Did you have feedback? What is that gaurav.

L

No, I was just asking that the second view is same as earlier right.

K

Yeah so like uh I wanted to ask one thing uh like why: why do we have control over the code? Mirror thing like to make the numbers go a little bit this side like it's, then the field is like overlapping. The numbers, like I didn't find any controls of the code mirror to do that. But if there is, I missed something.

H

Did you use the same code mirror that the actual.

B

H

One is using if you go there and click on it, and so uh then we pro probably uh margin or padding issue or similar. That is stopping this. uh You know this is rendered correctly here, and so you have to inspect the element and see like what's actually, you can also uh actually have also you know, control over code mirror too, so you will have to just see the code mirror docs.

K

Okay, okay, I'll do that yeah and yeah. I think I had one more question like how is the like this thing working like when you click it? It opens a configurator form right, like I, uh I didn't find in the code like what like on click. It should open a pattern form like.

H

Yes, yes, yes, so the file name is pattern configurator.js. So uh if you it's under components, so just control p, you know, find the file and it's pattern. Configurator.Js.

K

Okay pattern: config.js okay, noted.

K

Here and the deploy logo and the maximize button should work.

L

And in the card, if you want, you can also add the configurator button, so you can directly access the configurator from the.

K

Card view itself, yeah.

L

This is the configurator.

K

Button like this will be very okay.

I

If the size of cards are like, the cards are big, then like they are not the area is free in the cards.

K

I

They have big, maybe we can shrink them and get it three in line. Maybe.

K

Three, like three columns.

I

Yeah three columns. If if, if it looks better and like that's just.

K

I

K

Yeah I tried to like lower the uh this length of the card, but it didn't look quite good, so I increased the padding like margins or for the buttons to make it as big as the performance profiles, but I never tried to decrease the width of this.

C

So a piece building on that. So basically this is the first iteration of this. We basically want this to be sort of a catalog, so we have the white space here for now so that later on, we can show compatibility and some compatibility then caveats of having of using this pattern and how many times this pattern has been used. We want to show that information here as well, so just to get things started. This is the first iteration of it. So that's why we have the white space.

I

K

Okay, that's it! Thank you.

C

Thank you feels weird to thank yourself moving on I'll. Okay, we have some time uh ashish. Would you like to talk about the updates on android testing.

C

Okay, ashish is not on the call uh lee about docker, desktop and machine. Just a brief introduction uh introduction. We only have three minutes.

B

Yeah, it will I'll postpone it. We can have it as a future item um the same for the rest of the topics I actually just, maybe maybe time best spent.

B

If you all don't don't mind, um is to just kind of follow up on what aditya is what subedi is working on, and that is that in the um community uh design file for master ui, there's a page called patterns, and so there's sort of two proposed iterations on what these the grid view might come to look like and alvin has been helping quite a bit, so we'll have to um sue betty. You have to let him know, send him word of your progress.

B

I mean he'll, be quite interested to see his designs in action, so the first like to take a step back overall. The goal is essentially.

B

Is more or less to kind of reproduce um the service mesh catalog uh like in inside of meshring, to help make this a little more convenient? There's nice um discoverability that happens on a web-based interface like this, where people can come, find and pick. You know available content for, for you know any mesh out in the catalog, and then you know bring that we need to facilitate the import of a given pattern into meshri and we do right. You can go into mesh. We take that that url and import the pattern.

B

It would be equally as convenient if your starting point is from within meshri that you could more or less browse a similar card view, style, catalog, and so that's more or less. What you know. Civetti is iterating toward in the first iteration he's looking to take advantage of the existing information about the designs about the patterns and create kind of a car.

B

You know this card view this grid view style in which you can flip it over and kind of look at the details behind it um in this second and so without any other code change to mastery server. I believe that this can be accomplished, what's described here in the second iteration you. This might start to look a little bit more like what you've what we were just showing on mestre.io, um and this would require you know a couple of changes on in measuring server.

B

I think to get here yeah, so just for for context.

B

So we'll have to we'll have to send a word to so there's some other um other things for you all to consider, as you think, through this and what's useful to users is, it looks like alvin is proposing some filters. I I haven't seen these, but that's great um he's. Also we've discussed kind of like big numbers in the past and how you know some certain statistics are helpful.

B

Sometimes the present you know the presentation of those statistics in a line chart doesn't, you know really make sense or it doesn't really help, but but a big number, if you will like this or or even bigger, is um insightful. So if you have a particular design, um how many infrastructure components are represented within a given design, how complex is it is is or that that's one signal like how the order of complexity- I don't know if there are other big number designs at hand, but but food for thought, as you all think on this.

B

Any other comments on that anything else that we missed on today's call.

B

It's nice to see smark again uh on the call smark nice to see you. It's been a while.

C

Okay, uh so nobody wants to say anything else.

C

Okay, silence, okay, cool! So thank you all for joining uh since we're at that time. Right now, so uh I guess I'll see you guys next week. At the same time, thank you for joining again, uh see you guys later bye.

B