Layer5 Community, 24 Jan 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Layer5 Community Meeting - January 24th 2020

Description

Welcome new community members @ivanabyov, @Cybercynide, @chmazur, @flamenco_raul!

A

Good morning, well, that's such an awkward thing to say to to three people who are not having morning right now good morning. I have on a good morning or good evening.

A

So if, on a remind we're, gonna we're gonna chat a bit just before some others joined but remind me, am I getting my geography wrong is. It is Belarus.

A

Okay- okay, hopefully that wasn't this monster, got it okay, so I've just I just insulted you in that. In the first 90 seconds of you, people.

A

Nice I'm gonna grab the link to the meeting minutes and pop them in good morning. Dr. G I, just I just insulted Ivana in the first 64 seconds of her being on the call.

B

A

Sure I insulted someone for sure I'm, not sure boy, something.

A

Nice so I popped meeting minutes over in in the chat as you come in yeah. Please do record your attendance. We're gonna have today we'll probably have a couple of first-timers on the call, but but some folks, like Ivana, actually Ivana first time on the call but Ivana has been I, was actually already presented. Mastery, which is fantastic so.

A

So any road good to see you.

A

By the way it focuses as as we introduce mr. Jane here anywhere, you might have, you might hold the record for fastest time to submit a PR after joining the community. I'm, not sure I. Think Raghav is the last record holder there, but yours was: was it like two hours? Was it 90 minutes I'm, not sure it was fast.

A

It's fantastic so we're just going to prep some of the meeting minutes for the next couple of minutes, while some others jump on yeah. A PR is a PR right yeah. No, it's good.

A

Okay, if Ana's here so we're gonna, one of our topics is the an assignment after enhancements.

A

Dungeons with this very good.

A

A

Very good font cheese with this.

A

Hey Fonzie, quick question for you: are you with aunt della now permanently I, don't know if you're, in a place where you can talk and happen.

A

Funking, a Fonzie on that call.

A

So funny I, don't know if you can talk or not, but but in della has been I, don't know if everyone's familiar, but um okay, no okay- maybe it's not worth talking about that I. Just one of one of my colleagues is on the board. I recently joined and Ellis bored, and so anyway, anyway great we're five. After let's get going, we've got I'm gonna make everyone feel uncomfortable and and say: we've got Ivana's on the call for the first time, and so is any road.

A

Did you guys want to just take a brief moment to introduce yourselves to the rest of the folks I.

B

Am a bother and contributor to lady Sam.

B

And I I saw Sam measuring seven doctor if we invited me to, as their examples were able to sample those little ball recently seen in this bed. Examples and thinking that they will be able to integrate with measure in when someone asked you.

A

That's awesome, yeah, actually, speaking of record-setting Ivana, actually it's worth calling out that Ivana I don't know that anyone will ever beat. This Ivana has the record for fast quickest time to make a measuring presentation. I mean literally like five minutes: yeah yeah cube Khan, no less so I, don't know, if so I'm, pretty sure for sure. Well, I thought it was recorded. I guess.

C

I'm not I'm, not sure, but no, it was like.

B

Yeah yeah and that's.

A

But still, but still it was, it was pretty ridiculous. It's pretty pretty impressive! So for those on a college, just as you get to know Ivana's so over VMware has been monitoring the open source program. Oh no, not the open source program. Oh yeah, okay and yeah, yeah you'd, been I. Think yours, I, understand. You'd spent some time. Also I. Guess when I, when Alex Ellis with open fast, was there I mean? Did you spend some time on open files as well yeah.

C

Congratulations to.

A

You I just I've known known, alex for a long time. Some, the guy is the perpetual cheerleader for open fast, it's impressive anyway, and so, and so yeah Ivana holds the record for fastest time to present a mystery at school.

A

Any route did you want to do a quick introduction.

D

D

This is my second year and to attempt gzip and this time I'm trying to explore projects related to CN CF. As it's a very big project like CN CF is a very big foundation so that looking for good projects related to Cuban eighties and continued orchestration, just to learn more about system design and cloud infrastructure.

D

A

It makes the perfect sense.

D

A

Its way here- because this is this- is the project to be so other than other than network service. Mash.

D

It's taking really quite some time to actually understand what the project is about. Like I just sent you a message, yeah regarding.

A

What service is you know we if you would actually enter it, you would I know it would benefit everyone on the call.

A

Maybe if you take some of those questions pop them into the meeting minutes and we'll, let's see if we can't talk a lot about them, because because, like every other person, that of the other five 550 people that are on this this slack and have come onto this, call that that they've all very much so most people are new to cloud native. Their new two containers need a container orchestration new to distributed systems.

A

All these things and mesh, we as an area of focus rides, is sort of predicated upon all of that technology being in the environment and rides on top of essentially all of it, and so, if you can, for those that have been successful in sticking around and spending some time here, they've greatly benefited from starting at the top. If you will and and going down because they, because in focusing on service meshes, they then inherently come to understand all the rest. So.

D

Should I pose that on hashtag mystery, the same method which I sent here absolutely yeah, yeah yeah I'll get it done in a minute. Okay,.

A

Okay, very good, nice and greasy I, don't know if you want to I kinda I kind of hugged the the host there for a minute or if you want to take, maybe share the meeting. Man answer.

E

um Yeah I mean unfortunately, like you know. You know the fortunate that, like you know, the conversation has been going well, so I didn't want to interrupt. You know. So that's the part, like you know, say I don't want to jump in so I've already been making a messaging. You know on the side, you know it's, it's very nice to actually have all these new people join in like Nirvana Anirudh.

E

You know thanks Laura for joining and we have- or you know, regulars like Anand and Gunjan and kanji, and his news is actually becoming another regular attendee for this colleague in which is which is great, like I'm really happy like I. Don't have all of you guys here so other than that, like I'm, pretty sure like you know, we had some announcements but no, which he'll probably get to.

E

He did talk about like know a few things here and there, but one of the new or interesting things that happen in the last week is that I should go. He's he's not here in this call. He is because of his time zone. He. He actually took this initiative like an hour with ganjam and I. Think I forgot the other person's name. So so it's actually.

E

This trial, like you know who who took the initiative on creating help chart for ma cherie and now, since our Chico was actually much more experienced, looking on that he kind of nailed it down and he created a PR last week and that PR has landed. So so we we have a. We have an official chart.

E

You know for mystery and I think to of Theodore per second I. Don't think it's completed for all of you rappers, but I think he had completed it for is T on Liberty so but that'd be by yourself. It has been with a long time coming out and and yes absolutely, this is so. This is a big thing like now this week, which I'm very much excited about and I I see you I can probably like Anna Lee has actually added this year, so we have a person. His name is Dee Raj.

E

You know he was from Citrix, so he had company the CTX adapter and he had been working on that. Like I mean he ran into a lot of challenges, and you know we helped him get asked all that so now it looks like you know that a rapper is actually eventually making its way to becoming a stable or after waiting in the mesh we ecosystem. So that's that's phenomenal.

E

Now the third topic, of course again, oh, so this is mainly for Lana, but you know, since you're here you know just fantastic is the NSM hopper enhancements, but before we delve deep into that topic, because you know I'm pretty sure there will be some discussions around it. You know I just wanted to highlight, like you know the issues that were open and closed last week, you know so so altitude was a pretty busy week. You know theme has been like doing a wonderful job.

E

You know you know working on issues and crunching and closing down I'm, pretty sure. Like you know, there are many more folks like you know who were not able to join because of some some. You know festivities in India I think, like you, know festivities and troubles and some holidays in India, because otherwise like no, there would be more students from India likely would be joining like and and but but the good thing is, like you know, day definitely worked through the week, because many of the issues that were close were actually posed by them.

E

So definitely a great thing. The interesting thing is Shubham. Actually messaged me earlier that you know he is not going to make the call, but it's interesting to see him characterize. So it is very interesting alright. So you know if, if I, if we accidentally missed any issues, you know please feel free to actually add it to the list, and the other thing is: if there are any other topics like you know, you would like to talk about or bring it bring it for.

E

Discussion like you know, feel free to actually add it to the list of topics, so we can actually revisit it and young spend some more time in it. So so now I'll I'm going to leave it off here, like the ASEAN and, let's talk about, like you, know the NSM adapter enhancements we want to actually you know, work on which will probably be an interesting area for Iran to actually chime in and help us I.

E

I didn't start off with some simple things. Like you know, one good thing is that I think this is something like I, which we already reached Jana with now. Anacin is actually the first reduction where we actually attempted a helm, install of the sunfish. Then I brought those fish because all the other, a perfectly we have are not doing it using home, but we definitely have it, in other words, not to actually get back to it and like I, don't make a hell, minstrel work, so so NSM elaborate, but there is definitely an interesting one.

E

For my sri in general. Now one of the things is that as part of our as part of the in as part of the first initial Aniston adapter development, like we also try to bring in the sample applications of requests that were part of the NSF adapter rebuild so I. Think one of the first asks well, you know we had for Ivana was actually to help and word all the other sample applications which were kept separately.

E

You know and bring it into the adapter as a set of sample application, so people can actually try to try them out from mushroom. So I think that was the first ask we had a few on and you know, even if you want to speak about it, let me know please feel free to otherwise. You know.

B

There is them, and I first need to.

B

B

A

Actually, Ivana on that I don't know to what extent it helps to have kind of prior art if you will in but um the new.

B

Yeah I know he I think it's trivial, but it's bureaucracy, convinced he there are measure is in the list of projects that we contribute to and be able to open poetess from my totally.

A

Understood I'm having having established the open-source program office at SolarWinds, I. More than understand the bureau, the red tape that comes with yeah, okay, good good.

F

B

Mystery to discuss here, we can go with the other things, cool.

E

Sounds good, alright, any other leader you want to add anything to that is what opera enhancements before we have some jump into another topic or something yeah.

A

Maybe I'll just just share these share. My screen momentarily just to help Wow Ivana is waiting waiting for the okay and some paperwork. To finish, yeah I did you'd said. I came actually I think that one of the interesting things about the bringing in helm as the deployment vehicle for NSM is that one, it's nice that just to your point, to reach that escucho had had just landed the home chart from ESRI, which is a guest, is related, but we should remind me today: have we provisioned? Does it mess reprovision anything with using home, though.

E

Yeah, it's the NS on the wrapper. That's the only thing like we are provisioning with help. Okay, yeah.

A

Okay and then this would be bringing in all, hopefully all the rest of the sample- apps correct from there Sam, so the okay. So it's nice, then that the the some example code, if you will is already in the NS Emmett I,.

E

Mean the NSM inoperable, hands, I, think three or four sample applications.

E

Now those were made available as far a time, so we have them as part of my food, but not the rest because, like you know, when we reached reach out to the NSM team, you know we were told that there are more sample applications we can probably use, but those were not yet being converted to hell, because you know the Helmand conversion happen right before the NSM con like a week before or something so after that, like we have been, we haven't, you know, got back to you, know updating, DNS of an opera with more style applications.

E

We don't know if more has been added to their hell repo if it is like no, they just need to kind of figure it out and like me, or get into the sample applications list. Nice.

A

And near as I recollect ivana, do you recall actually uh I guess I guess maybe the appropriate thing to do is since, since we sort of jumped right into that particular topic, um do we, if only it was right around that time, like Reese said that those sample apps were there were home charts coming forth of what we had described in in the issue?

A

I think we've gotten fairly specific about particular areas in the NSM adapter, where, where code modifications might be best made and I suspect you probably not Ivana, you've probably not had much time to actually give it a go or give it a look. Have you gotten a chance to just run mesh read locally.

A

Ok, very good. That makes a ton of sense trying to find that issue to to bring it back. Oh No.

C

Yeah, it's too stale about that put.

B

The one face: okay and it needs a bit little change, because currently how how the deployments are wrong, they are specifically listed, so it needs some changes.

B

Thank you update the whole thing so that it will function more logical to make an exception.

A

Make sense make sense so that they're not explicitly named so it's a little less fragile.

A

Okay, fantastic well donate. Let me just give a couple of quick updates and that's just to say that um Rahul Amerada has joined in slack I, think yeah last night, um so she's we're fortunate to have her she's gonna go write up an article or two about massery, and so that's great. So if you see her in slack, say hi Rahul has also joined this week.

A

He is asking us to help him organize a kubernetes community days in El Salvador and also to come and speak and then there'll be a co-located event as well and so part of his goal. There is to help bring. You know, he's pushing an initiative to bring actually three thousand three thousand something I'm cellular base stations to El Salvador, to help make sure that you know to help get people connected to the Internet. He also wants to help get open-source going in El, Salvador, and so so.

A

Hey soos is on the call now and and that's part of hey Seuss's mission as well and and with his focus being you know more more within Mexico and so hopefully, we'll find hey. Suz will find a lot of synergies with Raul.

A

Chris also just joined yeah. You know, go ahead, aces.

G

Yeah I'm hoping to meet proud to tag with some in some Spanish.

H

A

Also Chris join this week. Chris had helped organize the first service mesh day before service mesh day was donated to the CN CF and is now called service mask on, and so I think Chris is I would expect that he's wanting to help us in the in the service mission community in layer, five he's probably looking to help run an event.

A

We've been long being asked to create a cloud native day, a cloud native con in Austin and so there's a couple of other cloud native ambassadors in Texas that are volunteering and asking to do that. And so so we will see also John dress joined this week so that that's great he's been really interested in contributing and he just introduced himself. We've got a vana. We've got an upcoming meetup at the cloud native Austin or initially my shirt will be presented. So that's good, shoe bombing was actually I.

A

Think getting to the point where Shubham did did. Were you able to finish all of the workshop previously and I suspect you might not be able to talk so I'll. Look for your chat, but just I'm excited that you might. We might finally do that with um Jaipur. That's good. Also, there's a new new sections on layer, 5, io Forrester had included a layer, 5 and measure e in a recent report that report and the webinar. The presentation that was given on that is in the community drive the measure a community drive.

A

Thank you, I, believe it's under presentations, there's the webinar and if you don't find the report, then just ping me because I'll make sure you guys get it, but those are the general announcements. So it's so fantastic.

A

We can go back and either I guess you know, since Dheeraj isn't here, the the the deal here is Saudi Raj was able to get the Citrix CPX adapter into a bait bait estate, and we need to give it and it's pretty much, probably ready for advancing to stable. It's just that. We need a few more people to beat it up, and so I'm gonna include a link to the like, so anyroad and Ivana and and shuperman, and everyone else that the I'm gonna include a link here. That.

A

Will give you the right mesh reconfiguration to be able to run the Citrix EPX adapter, because it's not included in the default mesh redeployment today, because it's not we have it as a gret that adapter hasn't graduated to stable, but, as you guys are poking around, it would be great if other people are taking a look at it and trying to use it to see if, if you find any functional defects and then so, we talked about NSM fantastic as we come down and think the reach there's probably a couple of different topics.

A

So Nuno is gonna present on Porter next week, both running measure II as a porter application and know does everyone, hopefully everyone's kind of familiar with Porter cm and docker app as they're. All somewhat related.

A

We've talked about this a couple of different times on the call, but I'll put a link to see nab that I. uh You know it has been helping us get measure II into so kind of. Similarly to how we got my cherie into a helm, tart we would measure would be available as a porter app and a docker app by the way. Just with docker captain hat on Oh, actually I'm I forget so next week measure you will be presented to all of the docker captains in.

A

A private forum- and so hopefully we'll see a couple of them joined here, I know one of them. Loop, jugglery was looking to join.

A

So sorry, so that's it! That's exciting! That's good! Going back to what I was saying. Oh.

A

Is that as docker Reese Wiggles, because I'm sure most of you know that docker, as a company had recently about half of it, had been acquired by Miranda Snee other half continued, the the other half the sort of the docker? As you know, it will continue forth and have much of a developer. Centric focus part of that focus will be on new use cases for docker app, which you know, which I think are exciting. It may be that so Anand it ananza I know it's on the call right.

A

It may be that yeah yeah, and on that you know today, measure II, CTL, wraps docker compose fairly prominently, but as enhancements come forth with docker app, it may make sense to begin to also leverage docker app. So so there's there's your next assignment. I guess begin to get familiar with docker app, because it's a bit more convenient and we may find Anand that you might find some reprieve because I know, as as users go to download.

A

Massery CTL running the measure containers requires two prerequisites on the on the local machine. One of them is docker and the other one is docker compose if the user has docker desktop. That's no pressed fine. That docker desktop includes both docker and docker, compose kind of docker engine and docker compose that isn't the case for those that are run, maybe you're running on Linux and have just installed docker and not necessarily the separate binary, docker compose and so on. On Dino you've had to go to additional lengths to do prerequisite checks for docker compose anyway.

A

My point is docker app when it emerged about a year and some time ago, a year and a half ago, it was as a separate binary as well and I haven't kept close tabs but I'm. But actually anyone can go. Try this right now, if you try to go use, docker, app, I, believe it's been incorporated into dr. engine itself, so so Anand in a good way, one less potentially one less prerequisite to have to manage.

A

So great, so all that's great of the half of the call. You know we should either anyroad. We can talk about some of the questions that you had brought up or if others have something to demo or if others have want to talk about some issues that they've closed or some open issues that they have questions on I'll go.

G

Ahead, I have a question about the the the blog is there lamp or four star: writing blog post I.

A

Answer that any more enthusiastic.

G

To be in other meeting is not ready with community or or is in development meeting or in which no.

A

This is highly appropriate for this one, as a community call yeah that so I think two things two comments there, that I don't think we made much noise about it, but one of our contributor, you know one of the community contributors Rafi he had he'd participate, he's still participating in the community, he's been more focused on the the other project, the service mesh landscape project and significantly enhancing.

A

It he's actually made a significant enhancement that it isn't hasn't been merged, yet it hasn't been released and Guri she's helping him there, but anyway, he's been writing about his experiences in the community and working on the landscape, and so we've begun to publish those community authored posts here and and yeah I mean that's in part, I think for um Radha. As she author writes something she'll likely have it published in on earth, uh I'm, not sure.

A

If she'll have it on open source, calm or she'll have it is she she's got a couple of other venues to post on, so it may show up in a couple of places but hey-zeus? Yes, let's, if you have do you have that particular topics or ideas that you're thinking might yeah.

G

I have a personal service facilitating preparedness.

D

A

Knew know who's presented on the call. Quite a bit. He's also dropped you, so he sees of what may I'm gonna point out separately, I'll kind of point out a couple of resources for you to help you get going in that regard. Yeah Nuno has got a couple of blog post drafted and he just has the hardest of time. I think finishing him off so you'll see you'll, see some stale ones in there as well.

E

A

Before we go through honest questions, any Ritz questions, anyone have other items to talk about.

A

There's I'm gonna talk about one only be only until the point that someone interrupts me because, because it's kind of neat, there's sort of two areas where again layer, five and measure e specifically are being called out so I I know. We've shown everyone this in the past that initially, when service measure interface was announced, that measure II was a partner there and so we've kind of appropriately updated that with the community logo.

A

And so so it's in good company here last week, I gave an update that SMI is under review for being incorporated into the CN CF and so the as I chair, the CF sig network.

A

That group this SMI group that we participate in brought SMI for review and so, if anyone's interested in that and reviewing that that either the recording or the presentation, I'm gonna, take a note. I'm gonna go ahead and find those links and put them into the meeting minutes.

A

But but what other topics do we have.

E

She actually didn't have a question about testing one of his PRS, but since he's not able to talk, I think that you know we might have to push it for next week or something.

A

And his guns, are you on the Cox I? Think you asked you ask a question recently and I I overlooked it in the channel, but I. Think part of your question was that you were looking into the you're looking at some liquid in Jekyll and wanting to understand kind of the I think how to iterate through a collection, she's, just a great class.

I

Didn't know I took the solution. Oh all,.

A

Right you're too fast, okay, I.

I

Have a question said as the the same chip or hinted for a machine is lady, so should I add the Machine him joining the Machine hub, sorry, game hub.

A

The accretion commenter.

E

We sure can but I think it'll be better if we stabilize the chart and then you go there. Okay yeah somebody else will have to approve it. Let me know, but you know we haven't, had a chance to completely test all the capabilities which are Chico's actually added, so I think, let's, let's we will get wet like now, but that's a great suggestion. I! Don't we that's? Definitely like you know in you know we did have it like in our future I'm, not by future. It's not like a far future.

E

It's like a very near future. So probably there hopefully like no once we get tested, people trust us to go and we have added all the existing the wrappers to it. Then you can actually take it to help hub.

A

Yeah, nice, that makes total sense, and so the short answer is yes. We do want to take it there, but to grease his point, we just need to run it through some diligence and I guess you know finish off the last bit of it too. So.

A

On actually as you're going and just getting steeped in cloud native and deploying missery in a couple of ways, you might want to give the new helm chart a run.

D

Can I get a link to where the shot is or is it in the current binary.

I

I

Happy to say something: please, oh.

A

Yeah I can't really, if you would keep, could you point anirudh to both the functional specification for the helm, chart that's been written and then we're also pointing to where it got merged in the main measure. Ii repo, it's under the installations folder.

I

Look if the focus is thin double link in the chat in the chat box.

A

Nice time since you were here seemed like there was an octoroon announcement this week. Yes, we.

J

Related open source project there concerning the risk associated with workload configuration in Bonetti, so one is our framework for determining what is the level of risk of coronary workload configuration and the other one is cube, scan, which is our tool to actually run in a coconut gesture, and can the workload then get a report on the risk level?

J

So far, we got a lot of interest and hopefully we will have certain having people contribute more. What they see as risk factor is in kubernetes workload, integration.

A

Totally I'm sure you did get quite a bit of interest.

A

Maybe can you, if it's of interest to me and maybe could you say more about your guys- is approached to the KCC SS I, just yeah and and you where the rules and the scores are being sourced from and how how closely they relate to CBS s.

F

J

Inspiration around the coring method and scoring formula and the factors into the scoring formula are taking from CBS n, so we're looking at the impact of the possible impact of each configuration and that ability how complex it is and how easy it is to be exploited. But that is very similar to C theater.

J

As far as where we're taking when we started with the GIS benchmark, which is probably most complete, set of best practices for securing kubernetes.

J

And we went and extended that, with some of the things we saw in in kubernetes deployments, we work with external advisors that gives us their input and and what we learned is that really everyone who worked professionally in kubernetes has some ideas around what could introduce risks. Obviously there are a few things that are obvious, like privileged containers are privileged part, introduces some risk sharing the host network in it, but beyond that are things that are less obvious, relating to host mounts and mounting of service account tokens and auerbach privileges.

J

So you really hope that more practitioners around world will contribute their own risk factors into that report. Oh.

A

Nice makes a lot of sense. Dude is this cubism of some of those risk factors or any of those based off of the CIS benchmark for kubernetes yeah.

J

Many of them do CS itself is releasing a new version, so they're now I think at one point four and they're releasing 1.5 for J's benchmark for kubernetes right this one, and if you go to chapter 5 of the benchmark, it all has to do with workload configuration so yes, a lot of the things that we are looking at are.

J

Taking progress, benchmark.

A

A

And then a chapter 5 chapter 5 workload.

A

No very good, ok I spent some time helping get out the the the docker CIS benchmark for the the 1 dot of I'm a few years ago, and so they're they're worthwhile I mean that those are good practices.

F

Yeah for sure some of them are our heart to follow in full.

J

That you actually sometimes do you need to villages and permission for your workload, but in general it's a good set of dialogue. Yes,.

A

Yeah sometimes the security is just so prohibitive. You can't give anything down.

J

That it's about managing risk, not eliminating risk yeah.

A

Well, so Anirudh had had a couple of questions about just grokking service meshes I, think in general, and we're fortunate that time is here both as he's the CTO for octarine and with grecian with Ivana and others here that are well steeped in service mesh anirudh. Do you want to sort of explore part of the questions that you? If no one else has anything I think on your roots? Questions are well placed and would probably benefit the rest of us.

D

I just wondering like: what's the use of service members, currently my projects I simply created a post like Sybil I I created, so this pose dress, I, see I created a back and I simply used to connect it using api's or some allowed port forwarding and and all the stuff, so I just wondering where this service mesh from into plate.

A

It's a question everyone has and actually, since I, get sick of listening to my own voice in this call. Okay, does anyone else want to give it a take a crack at that answer,.

J

So I'm not sure.

J

What is the exact architecture here, but there is a certain level of complexity which is kind of an entry point, I think for a service mesh. So if all you have is a database and service fronting that database.

J

Perhaps service mesh is a little early if the way I said it's more, when you have some internal traffic inside your cluster, that is either HTTP or grpe, based in addition to any TCP traffic and then service mesh is very useful both for traffic shaping or for monitoring and collecting telemetry, eventually for for security as well. But again, if it's a you know, one or two micro services fronting a database, I'm, not sure you'll get a lot of value from service much.

D

The thing is: I just give an example for a database and a so my main issue. Look generally for any service you create, you could just simply create a route or there's always a route and an exposed port which you can access to communicate with the service.

D

For example, if you have a 50-minute it gives in cluster IP that you can use to communicate within the cluster. So if even if I have like around 10 20 different pieces of services, that I could simply communicate between them using their specific IP addresses that are no difference for the DNS. The yeah 30.

A

D

Actually, other in the names here, yeah.

A

Oh add some flavor to what Heim had said and, and that is in part, I think some was saying is that service meshes are quite sophisticated and you know the different ones offer up different functionality, but in part what he was saying is like hey, the more sophisticated, your environment and it is, and the more needs that the that you have the more services you have, the more complex your environment, the more value that you'll derive out of the level of the out of the capabilities of a mesh, because there are other tools present today that can get you from point A to B or kin that you might not need all the power of a service mesh initially in in the type of example that you were giving and actually recognizing that you were giving.

A

That example purely just for illustrative purposes that to speak to that a little more directly. If you have a front end, if you have a front end application, and it relies on a back end database that, given that you know I, don't know that people would always refer to that as a distributed system. But but irrespective of that term, if you've got a network connection between two components of your application, the network is fallible, it could break at any point. Your database could be rescheduled and moved around to different nodes or different somewhere else.

A

To to your point about, like IP addresses or so service discovery is necessary. You get some service discovery out of any container Orchestrator, and so, while you might not have your your front end app hitting your database via by a named IP address, it might be using DNS, it might be using.

A

You know just just the hostname of that container and so even though the container moves around your, your front-end application is able to track where it's going, because it's it's it's just calling it by name and when it does that look up it does the DNS, lookup and kubernetes is running its own, managed DNS, and so, when kubernetes reschedules that database onto a different node, it updates its own dns and tells you front-end app to go so a lot of that is kind of taken care of.

A

But what isn't necessarily taken care of or any number of other considerations, and that's that maybe the database is performing slowly or maybe it was being moved right when you were trying to call it, and so you might need to retry. You might want to have your front-end retry that connection to the database when it get moved or maybe when it's running slowly.

A

You want to again either retry or back off and wait for a while or send your that, your client, your user, a notification that says you know that that gives them a better experience, telling them that we're processing your request and the service mesh or almost all, not all of them, but most of the service meshes, are capable of facilitating those automatic retries the backing off, and we trying again that the the redirecting traffic either permanently or temporarily, and it gives people a lot more fine-grained control over how requests the resiliency around service requests between users, calling any of your services, and so Hinds Point is like hey the more complex your environment, the more services you have like the more dire your need for a service mesh, because you don't want to be coding.

A

All of that logic into your apps. You just want to focus on what your app is supposed to do, and you want to Specht more intelligence out of your infrastructure. You want to just be able to rely on your infrastructure. To you know, do the right thing and so surface meshes are quite configurable. With respect to that example, that I was giving about like retrying the connection. Should you try 10 retries, or is that gonna give you you're?

A

You know you want it to be resilient, so you want it to retry at least once, but if you try to eat right n times, is that appropriate or inappropriate for your user to wait that long, and sometimes it is sometimes it isn't- depends on anyway. There's a law there's a long list of functionality that that some of the more capable meshes bring and it has to do with enhanced observability and the telemetry around it.

A

It has to do with being able to set enforce policies based on the fact that you know, for example, that maybe Anna Ruud was signed into the app and actually he's called the database too many times more times than he's allowed to within that hour, and so you know he hits a rate limit and- and you don't have to write that type of capability into your code any longer so long as you're just telling the service mesh hey, look for open up this cookie.

A

If you see Anna root signed in and he's called, you know the backend more times than we've defined that he that he's allowed to then block it. So it's that type of a thing that you can just configure your infrastructure to do.

D

In brief I guess it's more but fall backs and metrics and.

A

Functionality yep- those are two very extraordinary brief examples of yeah you're right, you're tracking correctly, there's a and the list goes on, I guess in terms of an identity and encryption.

A

Accounting and visibility.

A

I'm will talk to you for a couple of hours about security and how you can how you can run your infrastructure much more in a much more secure way, predicated upon having a surface mash in the environment and so.

D

I said, that's very like it cleared out most of my questions. I had about using service meshes and why not simply use the current trend. Yeah.

A

It's a good yeah in part. What time was alluding to is just that that okay was that.

A

Like if you're, if you're deploying services into a kubernetes environment, that you'll you'll find that many of your commonplace concerns, and maybe some of your lower-level concerns are some of them are addressed.

A

Some of them are addressed good enough, but if you've- but you might find later on that as as you or your organization or your environment, matures in your knowledge and understanding of all the ways in which you're exposed from a security perspective or or your environment has grown to many more containers and many more services, and you really got time to lovingly care for that. You just you end up really needing additional tooling beyond the service discovery that you get from kubernetes are beyond the network policy. That's capable there I think.

A

There's there's much like much like writing a piece of software there's a lot of ways to run your infrastructure, and you know, as a technology industry, we there are moments in times where we converge on a particular technology and say this is kind of the canonical way. The generally accepted practice in the industry for running this type of these types of workloads, and that for this genre, for this era of how that's done it sort of started with containers, moved into orchestration and is now you know, moving into surface mesh land and takes people.

A

It takes people a while. Both the operators at the infrastructure to just move forward takes the organization's a while to move forward. It takes some bureaucracy and red tape a while to get done and sometimes- and it also takes developers as they're accustomed to maybe being responsible for the resiliency of their app or the encryption and the of traffic to their app is sending out that as they it takes them a while to begin to trust that their infrastructure is smarter than it used to be and to stop doing that.

A

But it's absolutely my belief that service meshes will be, as you know, nearly as ubiquitous as kubernetes is because why not you get all that for free? Why would you it's the same value proposition, a very, very similar value proposition as to why it is that people are writing functions and running using serverless? It's just that this value proposition service meshes applies to not not just short-lived event-driven functions, but the other 80% of the world's workloads, which are actually long-lived processes, always warm.

A

Services that are running and so yeah I think that we'll find there they'll be a fairly ubiquitous.

A

Well, look at that. I can talk us right through the end of the call and that.

D

Actually, it was new for me I, like hearing what so is measured. You know the last time I said: I first started with docker, then I heard about kubernetes, then I jump to Cuban. It is then now that now I just heard about service meshes. It's pretty new to me, like I, never look and hear what a few dogs and maybe look into.

A

Yeah and then I'm talking, you know or ping hime he'll talk your ear off about service as well, so really any of the folks on the call they're. All do you know it's a great question. I read anyone have anything else before we sign off for the day.

A

At gireesh we're good, you miss anything I.

E

Think really good, like I, don't have anything specific. It was good call, looking on very interesting discussions and also educational or fun people I believe you know a great way to end Friday about young people. Haters.

K

Gonna say in my adult: yeah: hey runner, Hayley, sorry I would leave shame.

E

No problem you just have to start off with tempo shuts and then let's get back yeah I'm just missing with you. Yes, what's up yeah.

K

So let me show you waste.

A

So for everyone else I, you know thank thank you for coming. If you can hang out, you know, please do, but we totally understand if you need to jump out, because it is the top of the hour.

E

See you guys next week.

K

So I was looking through that performance shaking speech. Yes in the settings, page I did the front-end part and when I summit some just do their own testing. It goes to the background page here so.

K

In the way, she's sitting performance page and when I said the chain, the parameters, the concurrent requests and the Curie's per second, it goes to the back end here: API, a dot log touched low, taste and preference. This is the back end here, so it is receiving the back end. Fine I, take it checked it with my printing, the things now then, I received the things in the measure performance by the store it pops up.

K

Here so Gary's told me to basically load the data from the shore. This one.

E

Rocco, can you scroll further down? Oh yeah, yeah, yeah, I'm, sorry, yeah I see it there, the so. So what you were interested in is actually not the load test itself, but the Lotus crafts. That's the object which is going to contain, like in all the variables like now, you're interested in yeah.

K

E

So technically the way you have to do it is you will have something so you're submitting to the backend. But yes, we shouldn't stop with that, so you have to submit to the backend. Now when you get back the success from the back, you have to actually you know call one of the action action methods actually in the store file. Yeah.

K

This right here.

E

You wouldn't be working that, but can you come down.

E

Yeah, so you see the line number 166 onwards. Those are there yeah. Those are the action methods, so I created one or you know the load test, press yeah so line one night, yep, hey yeah, it's the one you need to invoke from your code like you know, which is after I mean you can find this pattern in all the other components like, for example, in the one you were looking at the performance component or the settings indicator.

E

So so this pattern is followed in all the places where a college maple back in now on success. We actually update your local store as well. So so that's the pattern, you'll be following okay. Now, without you updating the store, you will not see your changes reflect properly. So final, like.

K

This right, one second.

K

Here, basically, this is what you are referring to rate all.

E

Right, yeah, exactly that's exactly I mean so you'll, be doing something similar for your function, which is going to be like I, do know what I have yeah, so you will be doing this exact identical thing for that. So what this will do is this will not just update you know the back end, but also update this for state, so that you know they are in sync with one another.

K

Okay, and so basically thing having a doubt with, is here in the load tested and the load is here itself the resistant static value that is given permanently so when I leave it empty.

K

So when I lead I leave it empty, it doesn't mean that the the static value that I'm giving is will be empty.

E

Yeah you're right: no, you can actually put thirty years there. The reason why you need is empty. There is that now, if you actually go back to your browser and then load the page you'll see something you go back, you browser should.

K

I write 30s before.

E

No I leave it for now. Thank you. So now can you go to the network tab.

E

Okay, can you reload the page yeah yeah, just reload the page so now, when you search for our sink.

E

Yeah exactly now, how can you go to the preview and expand the data there? If you look there, you see the load test, press yeah yeah, so so the point is that you know this. Usually when it starts, you know it would usually like it'll return back this. This data, which is kind of getting stored in that in that you know, JavaScript, object, yeah.

E

We do the reason why, like you know, it's usually I mean it might not be a problem because it's probably coming from the back end, but if it is not, then then probably what are you doing? What you're suggesting is right, and that is if at all like no it's not coming right, then you should probably fix it of a back end, because you know well in the page loads.

E

This sink all it's just the current state of all the Preferences in the back end and then brings it to the front end and synchronizes it into the store. Now from then on, you just work with your store in the UI, because it's a single page application.

K

E

Yeah, let me know you know if you wanna clean any more issues- shell, let's take it from there yeah, but yeah. This is great. I mean like I mean one thing I would say is you know, don't hesitate to ask questions, you know what are you doing is right. You know this is great. Like you know, so everybody I think you know. If other people are on the phone, they would have probably learned a lot from this, and also it's educating you, you know.

E

So this is fantastic also, you know it's also a good thing too. You know it's, it's not bad to question something that's already there right, because if there is a better way to do something, you know I mean it's definitely good to talk about it. I'm gonna take it forward. So this is prayer. Okay, thanks.

E

Anything else guys before we call it a day and jump right into four weekends.

E

All right, thanks, see you guys next week, oh very.

A

E

A

Degree bye guys.