---
title: "Composable without Compromise w/ Natalia Venditto"
description: "Kicking off our first Composable without Compromise livestream w/ Natalia Venditto. In these live streams we talk about composability. From architectures, to design approaches, to tech organization and governance. Anything MACH Alliance, DXC or headless composition related is a valid subject."
date: "2022-09-16T05:04:56.000Z"
url: "https://timbenniks.dev/videos/live-uniform/007-7-ebcqip9ec"
youtube_url: "https://www.youtube.com/watch?v=7-eBCQiP9Ec"
playlist: "live-uniform"
duration: "1:00:18"
image: "https://img.youtube.com/vi/7-eBCQiP9Ec/hqdefault.jpg"
tags: ["composable-architecture", "ai-engineering", "performance", "cloud-infra", "frontend"]
---

# Composable without Compromise w/ Natalia Venditto

Kicking off our first Composable without Compromise livestream w/ Natalia Venditto. In these live streams we talk about composability. From architectures, to design approaches, to tech organization and governance. Anything MACH Alliance, DXC or headless composition related is a valid subject.

Watch on YouTube: https://www.youtube.com/watch?v=7-eBCQiP9Ec

## Transcript

yes [Music] [Music] [Music] all right we are live natalia what's up how are you i'm so excited to be here yeah me too this is like our first live stream that we do for something we call composable without compromise and so we already have two people watching and i'm really excited so um to get this started um maybe say your name it's a so it's supposed to be said in spanish because i will butcher it so i'll let you say your name properly and can you then also explain what you do and where you work things like that okay so i'm natalia because that's an italian name actually exactly i was just thinking and but i was born in uruguay in south america and now i live in barcelona and i work for microsoft and basically my role is program manager which in microsoft is a like the role you give to a role that does everything or well a bunch of things that are difficult to describe or to give a name to and basically i work in the developer division and what we do is we do developer tools we we own vs codes for example all the vs code extensions and then we own the clis that connect code to cloud and we also obviously because it's azure we own a few services that frontend developers or javascript developers can deploy their javascript too and what i do as a program manager i own javascript in azure so i make sure that the developer experience for anyone working with javascript in azure doesn't matter what what service even uh even in in intersection with m365 and the linux distribution of microsoft marinara oh the wsl of course um so i oversee the developer experience and write the strategy that we implement to make sure that everyone is happy wow that is something that's a lot and also you're kind of our go-to if we want something i guess right yes if you want something yeah exactly in azure functions container apps static web apps app service those are like the my main focus also cosmos tv and well you have to talk to me there is no way i found that i love it so in the context of this stream what we what we want to be doing in this series is i also want to go a little bit about the industry news so how about i'll share that now and maybe um if you have any comments on it i would love to hear your opinion also so here we go and so basically in the industry news of like the last month kind of i decided not to take too many things because there's so much going on but this one i saw yesterday that algolia actually bought search.io and that's pretty interesting because algolia was doing pretty well and now they also got searched io because they have something called neural search which sounds very magical how does this translate to azure search do you think that translates to cognitive services sdks that we can use to enable artificial intelligence in search capabilities so we do have a compatible analysis so there you go i'm thinking every news item i have you have something like oh yeah but we do that as well and by the way i also work with this case and community oh that's cool nice um and so then we also saw that gartner which is basically in the space of where we operate they are kind of like they do lots of research and they figure out what are the trends how things work and they released something called the hype cycle report last month and there's lots of things in there where they say okay the hype that is going on now there are like four or five of those and the interesting thing is if you look at where uniform sits and also what we're talking about today about the composability stuff it's one of the biggest hypes it's actually going from the model it's to a composable whatever composable code to composable front-end to composable architectures and so it's kind of cool that we at uniform kind of are figuring out how this all works right as gartner is actually figuring out that it exists and so that's absolutely it's a cool place to be and we just have to be very fast so everybody starts to understand what this means right and then yet another one a cloud 100 have you ever heard of that actually it's new it turns out forbes looks at like cloud native companies sas companies and figures out okay who has the most impact or has the most innovative services things like that and i put this in here because uniform has a whole bunch of partners and i think almost all of them are in here and so why not share this with them because it's not an easy thing to get in there because there's a lot of sas companies yeah and i think that was very necessary because we were always looking at the um 100 in enterprise or 500 in enterprise but uh cloud is a big part of the business ecosystem and we should really give some or highlight the work that is done there exactly and i i can bet that a lot of these companies use the azure cloud for these things absolutely i'm very certain of it yeah exactly and then this one just happened and when i wrote this slide it wasn't fully sure yet if adobe was buying figma which is like an amazing creative tool that basically we use all the time and actually now it's actually real it's true so they spent 20 billion wow that is it's amazing i love it i think it's worth it oh yes amazing interestingly there was a few articles not written by adobe themselves or by figma that were talking about you know the go-to market of adobe this year and in these first few quarters they actually dropped a lot of revenue like a third of what they had last year because everybody was starting to use tools like figma and envision and other things i personally use it all the time all the time yeah one of my my go-to tools right and what i love about it like i'm not really a designer i can do it but it's not my job and i still use it a lot and so they kind of normalize design tools and so it's very smart of adobe to buy them if people are happy i don't think so but we'll have to wait out how long it takes before adobe is on the logo i always i always say we need to give the chance to demonstrate what they want to do with it exactly yeah i like that and i think it's it's not photoshop it's just different and i think figma if they're smart they get all these amazing people at adobe that work in all these teams to get that knowledge and put photoshop like features and ai stuff inside the tool and then they win so by the way my mentor as an architect because right now i work as a program manager but that encompasses being an architect being an engineer and my my mentor for a great part of my career he's now a cto at adobe he's not there but he's cto for one of their um departments it must be amazing to work there because imagine doing the web version of illustrator or something like that it must be intense to build that very blowing yeah exactly let's not hang too long because i want to hear what you're going to talk about soon so i want to also talk about a little bit what we do at uniform not the whole thing because that's a lot of talking but also there's a few new things right and so just recently that's actually like two three weeks ago we have officially now released the fact that we work with different versions of edge functionalities of middlewares and things like that and we're secretly really hoping and waiting for azure to use these things as well um and so i know i'm sorry but like our customers are asking and we're now just diverting a bit and do it here and it works everywhere really really well and so we're excited to announce this officially now that we have the akamai one cloudflare and for cell and netlify and soon asher i imagine um and then this one is close to my heart where we this is a little bit older already um where we actually have our whole sdk working in nox3 as well including all the fancy new stuff of the nitro engine right so it can go to this to all the clouds and all the you know the edge functionalities and i think when we were both on a conference in barcelona you saw me fail at it because it was like pre-alpha i think you succeeded it you did a great demo i know where i failed when i went to netlify and that crashed there it's probably human error though so anyways let's not dwell on that part it was a fun conference to me because i met you there yeah same and and i want to say that we're going to talk about distributed systems and microprocessing usability and i think that one of the challenges we're going to see is debugging systems that are that are living on different execution runtimes and across clouds so yeah the perfect bridge how about that yeah and these are the ideas and tools for new cloud architectures but i'm not here to present on my own i'm here to have a conversation with you because you have similar backgrounds like i i was an architect before i i was an engineer before a front-end developer for many years many years too many and and yes we were chatting and that was amazing because we were chatting and we realized that we had the same pains and we had the same challenges and also the same joys we're not going to talk only about the negative stuff if you work in the thing that we worked in we know what it means to do something at scale and if it succeeds it has an insanely huge impact right so it can be amazing to work at that skill it's also challenging but that's what you're going to talk about a bit today right how you can solve that stuff absolutely so i'm not going to re-introduce myself because i already did um but we're going to start talking about monolith so you i i'm sure team you remember those big blocks yes big blocks rocks mountains of code and capabilities and yeah we basically had it all together like everything was a big application like consolidated yes in usually a single repository and had like massive teams working on it and then we had to come up with all those complicated release cycles and yeah integration models back in that that time or back at the time i don't think that we considered ci cd to be very efficient for us because oh no i've had crazy things like where i bought our boss walked into the room it's like why are you playing foosball or uh metegol in your language and um yeah i had argentine in uh colic it's almost how you would say it but they're like why are you playing that and he's like oh yeah sitecore is is building and the release is happening on ci cds like oh it's fine i'll see you in an hour that's absolutely that was completely normal and yeah we're talking two years ago here that's not that long ago yeah that when when you have like a compiled language and in the back end typically and you have those massive applications it can take anything from it can take four hours to compile it oh yeah seriously it's not a joke so obviously right now that's not acceptable for the kind of uh dynamics we have at the or the development cycles mapped to agile methodologies etc but that was how it was back in the day and it worked and then you had this layered architecture composing this monolith and the infrastructure was owned by a team always the system engineers i don't think we were back in the day talking about devops that wasn't a thing yet we had the system engineers and then we had people that were very skilled with databases and they owned the databases the database administrators and then the backend developers on the data access and the and the business logic and the front end was like an afterthought like in the it was like oh we we're going to make things pretty and move pixels around and that's that was basically the idea many um many of our colleagues in other with other concerns had of us um but basically we were very very much limited by whatever architecture and methodologies decisions were made by the by by anybody in those layers i think the whole thing right because i remember we would have to do like a quick fix for some css because people couldn't buy click on the buy button or something and we would actually fix it but then not release it because one release would sometimes for this project take 24 hours because all the systems that are so big there was multiple monoliths that had to sync kind of and so we would wait for a week to release it so there were more front-end fixes in the chunk but then it means there was a week of being people not clicking on the buy button it's like yeah that was kind of a normal thing yeah and and when we had something like a release model implementing a hotfix for example that meant and i don't know if that's in the next slide because um yeah i i haven't seen this this slides for a while but yeah we meant we had internal dependencies knowledge silos slow time to value technical debt and like lack of innovation all bound to this and i was i was um going to talk about the hotfixes for example a hotfix means that you have to bump the version of the whole monolith yeah so you you cannot just introduce some code just like that you need to release a new version of the whole application with the consequence if you've ever done it before that when you are going to the live server to the compiled code that is hosted in your monolith and do the change there have you ever done that before absolutely oh we have done scary like that you you have worked in adobe experience manager like me yes right and so adobe experience manager is um basically a java uh content repository so it doesn't have a database it has nodes that that map to java classes sometimes to to content notes and what we used to do is we went to the crx improv and change whatever code or value or whatever we needed to change because that was the only way we didn't have to wait for a new release or to integrate and then we would download the package from the felix console and name it after the bag or jira ticket and then we had to remember for the next release that we needed to include that code because we could cause a regression if we forgot about it so those were the um not so happy times yeah exactly but then if it works this stuff is quite stable it's not that fast but it works yeah and also the customer were happy that we didn't have to wait so long too because you say sometimes it costs money um if you cannot click on the buy button then you are preventing their end customers to have a good experience and do what they have to do so yes it's important to remember these interdependencies as well like if you have a huge monolith a change you may be doing here may be impacting in places you didn't even imagine and then the knowledge side loss because it's so large then we have people that only know about one feature or another and you have to wait for them to give you i don't know clues on what may be wrong so it's it's really complicated it's sometimes the right architecture for some kind of applications that's something that we might also agree with or admit but it's really really complex and yeah not for for the faint of heart really exactly and yes and and during this conference in barcelona i explained that in 2019 i was working with adobe experience manager we had this big monolith and i wanted to use modern frontends for a while already like i had i had applications right we had several different applications that were embedded like the whole application the whole sba in one page of other experiments experience manager but i wanted to have something more decoupled and yes i was a technical lead at the time um obviously i had to have the the crucial conversation with the architects the back-end architects i want to use microfrontends right and some people didn't even know what that was like maybe like i feel like sometimes micro front ends is explained in different ways how would you explain it is it like islands architecture that is now a thing suddenly or is it different i think island's architecture is a flavor of microfountains but it's not the definition of microphones it's a very good implementation of microfrontence but not the the only one yeah exactly that's more like a horizontal kind of um distribution of of components but you can have vertical ones like you go to one url and you are bootstrapping a whole application and that may be a micro front end the whole application thing and what i'm always worried about with these things is like okay i need to go to let's say as a user i have to go to four different things to do whatever in an e-commerce something complex so you might have four micro front ends and i'm gonna load the whole front-end bundle for that micro front-end every time and then if this is slow it's a bad experience for me and how is the state maintained between these micro front ends and those things make it pretty hard right how would you solve that yeah absolutely that state management and identity all those things are the most complex problems that you have to solve when you are implementing a micro micro whatever micro architecture yeah anything yeah micro asterisk architecture right and yeah it comes with other challenges like versioning like knowledge ramp like what are you gonna version because we said like yeah if you if you hot fix or you do a new release of a monolith you have to bump the version right yeah when you um when you make changes to a part of you like one microfrontend do you just bump the version for that microfront and would that be using dependencies like if you're using module federation for example and you are using and you're using shared uh scope are you going to be uh how you solve that like you you you may have to pump the version for a dependency as well that is a dependency of another microphone than right so then the knowledge ramp again the state management the data architecture because this is where you have to think oh one database for the whole microfrontend ecosystem or system or one database per microphone then you have governance and obviously orchestration how you put all together and make it work and yeah i think my take is that the first thing you need to to look into is governance like that's the key to succeed but anyway when i went and and talked to my architects back in the day they told me oh you're into my friends i love your slide yeah that's that's what they tell you like oh that's hype no we don't do this we have driven development a little bit um but there are so you have to come and convince them that there are advantages and i always say like the advantages are not only for the technology or in terms of technology it's also in terms of people right that have their lives enhanced by this kind of decoupling and that they can specialize and they can work in smaller teams and they can work faster and more dynamically and obviously the processes may be simplified because you don't have to worry about such a huge application anymore and you can also have different processes that are suitable for each micro so what's actually interesting about that fact about this governance related thing with the people i never considered this until let's say six months ago to me it was okay you go to these new technologies you're gonna make a better front-end website right your whole thing is better but that's not only the thing you're gonna win with you can actually go to all this new technology and build a crappy website but the internal governance is amazing because we used to see all the time with the the time of the monolith or the systems that were not nicely connected when your boss says oh i want a wish list for this christmas and you might be like um maybe in six months we might be able to do it but then when you have all these nice composed architecture things together in smaller chunks suddenly it's like you want it this week or next because i can just find the thing and connect it right and you can do the same thing in code idea and so actually i t teams in these big sis these big companies are now lovable again rather than oh crap this guy is just not going to be able to do it for us and that's a huge difference in business and then it helps that whole governance thing you can now i describe this as a relatively high level architecture choice right but you can do the same with code and smaller teams and be more agile between these teams absolutely absolutely they're they're the interaction between teams um is one thing the interaction between the members of the team um everything changes in this model and you have to account for it and and by the way those are the greatest advantages right you you enable those independent teams to make their own decisions you facilitate the scalability process or guarantee it yeah you make maintenance easier because everyone is focused on their tiny part of course they have to think about how we're gonna orchestrate this but if something breaks in your own micro service or microphone then it's just something small too or simple to simple now simple it's a bad one smaller a smaller problem and a smaller challenge and yeah you can distribute the workloads again like i was mentioning maybe for this micro front end the right database because it's i don't know it's a it's a scoreboard so let's have time-based databases behind it and maybe for this is a blog so we can have document model and then you have other i don't know types of content that can map to a different model of database or even skills in the team like you have a a group of very skilled sql developers you can have a database that maps to those uh skills as well it makes it very interesting i had a friend a while back a few years he went to a new company he became the cto and he he started to do this for his architecture but it was very new back then and he actually hired python developers he hired php developers front and node people they would make their own microservices and small apps in their team the way they did it the best they just had a little contract that said this is the way of the output and that's your kpi for performance nowadays you would maybe say use graphql so the contract is easier back then that wasn't the thing right but it's very cool that you have that flexibility right if you have star players in c-sharp why not utilize them for something like this absolutely absolutely and i think that i don't know if i speak about this later well this is about the hotfix because we kind of talked about it before but um yeah i i i lost track of yeah i went to that slide and lost track uh but we just got to the next it's all good yeah we go to the next it doesn't it doesn't no i was going to say yeah now i i know what i was going to say when you are distributing your or when you are decoupling your monolith and working in a distributed system it's better to adopt a kind of architecture that is api first for that same reason right because you don't you shouldn't have to worry about the api anymore you should have like something like a gateway in the middle or a management capability and and then it doesn't really matter what is the database in the end you have an api contract and your applications are all going to talk to these apis so yeah we already discussed this but again my challenge when speaking to these architects because they were in and that was the case usually until i don't know a few years ago most architects had a um a back-end background right so so when they thought about decoupling or splitting a part of the application they usually like let's imagine an application is it's it's an object it's something that you can look at they typically um if this is the the front end and this is the back end they typically would look at it from the back end so how can i oh yeah i'm just putting you back so you can show the object you were showing oh yeah well i i i'm not showing anymore i just wanted oh okay yeah we go like this we're good oh there you go oh angular there we go so you can um yeah we we we used to have them think about the coupling from this perspective right from the back end so they thought about decoupling into microservices and the monolithic front end continued to be part of the architecture in that way and then that worked as well because if you have a good component system and you have atomic components you can also compose a monolithic front end and it works but the build times will be um longer and also again we we're having the same issues you change like you you have to really have a good strategy and a good style guide for example to make sure that your front end works good when it's still a very large uh set of components that you're serving and so um this is this is basically how it worked like you have the monolithic backend you had monolithic front-end and they split it into microservices still with that monolithic front-end on top exactly and what is this i don't know because maybe i had some some clicking thing so sorry about the slides oh it's so good so and we are also going to split up the front end and that's kind of what this looks like right yeah so this this is what happens when the decoupling is done by front-end architects right you are thinking more about the user in the end about the user experience about the ui right i hope because i still see a lot of debate sometimes on twitter about wanting to adopt the technology because the technology is amazing for the developers yes yeah and i work in developer experience i agree that we we should have a good time when we are working because it's a it's a stressful job it's a yeah it's it's very uh it has a lot of cognitive overload and the more you can reduce that cognitive overload and make things simpler and the better right but the user the end user should always be central to everything and have more importance and more like be more priority than anything so when when from the architects couple of months they think about the user and they and this is this was basically the origin of microfront what is more effective optimal performance etc for an end user in a particular case and yeah what we want to do is go from achieving balance to achieving independence and every use case deserves its own architecture and tech app so you have a use case and that that's basically the user story as well what do i need to do what is it the problem i need to solve in this case from for the business and from the user and then make architecture decisions in a nice architecture matrix tech definitions have a whole development cycle defined and explore different building deployment options depending on the use case it's not one fits all anymore no exactly everything is very specific to a use case and therefore it can work really well from governance to a front end to an end user right absolutely yeah and this is basically what happened when we were working in those monolithic applications that we had a when we had when we look at the front-end setup we had a common set of components uh we had the html the html engine or render engine or whatever um and even if you had different page types that did different things like a very dynamic homepage you had a landing page that was changing all the time or or lived only for a while you had a blog page that was always the same or you had a user page that was very very um state bound everything was served by the same javascript and html and whatever styles loaded what we need to aim at right now is to have this kind of structure for our right so depending again on page type and the problem we want to solve for the end user the architecture and tech definitions that we're gonna have enable that solution so yeah basically that so if you look at this do you think it's harder for developers to understand or maybe it's only harder because they're used to the old thing how do you think i think that the jamstack has done a great work to enable architecture thinking in developers in front of developers like before we weren't very aware of what were the mechanisms that that make our sites work and until you were at a certain stage in your career right you were very focused in the component in the ui in the styling i think right now because of all the technological innovations and the emerging technologies and patterns and architectures we are more aware of this i am not sure everyone understands what should be the right architecture for their use case but this is something that yeah typically developer advocates and and and other folks yeah we kind of have to explain this because what i love about this now when you look at next and next and to a certain degree the new kids on the block like astro and sveltkit and stuff what they do is like their base is really solid so if you don't understand anything you build a website with it you're like 80 there it's actually 80 of an amazing project out of the box and so you don't even have to think too much about hey which bundles am i going to output just for my blog page like in the new nox3 instance for example you literally just it's it's smart enough to output bundles specifically for certain things that it deals with and it's always a statically generated site no matter what you do but it dynamically goes on top and then it mixes it up completely depending on where you host it and you don't even have to know anymore but then when you want to make the blog ssg and something else ssr you can just make that choice and do it yeah so it almost feels like people my age have too much knowledge you don't even need to know anything anymore you just go for it like isil has all these ie6 stuff in my brain that i have to understand and there's no need for it anymore right you know that's i'm happy i have it because i've been doing this for a while to see how cool this is but if you just are new it feels normal but it's not normal it's amazing yeah yeah definitely definitely i think it's going to come um a time when when developers don't or front-end developer full stack or jump stack how you want to call them um they don't need to think about this anymore either because well my my opinion my humble opinion is that um we're gonna move again to [Music] a consolidation or standardization of patterns like we had before like right now we have different ways of doing things exactly um but it's going to come to a point when we probably are going to have only the last two right in the front probably and something in between like if you want to like if there's a small change in your ssg stack of one page out of 20 it will quickly render just that one like kind of on demand kind of thing but it's it's going to be around those and the ssr bit i think is going to go to the edge there will not be any origin servers anymore absolutely i think that's that's absolutely the future and then you would have probably microservice um microservices containerized microservices that serve multiple um applications again through an api contract or an api standard but it's going to be to serve massive kind of um very distributed applications like i don't know twitter or something like this right so then you you need more power because you're doing a lot more data processing and computing but for smaller to medium applications i think we're going to be doing this data request on the fly render on the fly composability on the fly and and yeah caching whatever is um possible obviously but many things are going to move uh to the edge and not only the edge of the network the edge of the client like we're gonna have interoperability figured out between devices and and the in the network and the devices have surface workers already that you can use with yeah definitely the next three the new runtime it runs locally on my browser if i want to which is absolutely absolutely i think i am more at a broader scale and more standardized like more of an industry standard and and yes and this is when we are we're also going to be constantly i think porting and or swapping data processing or data computing between the edge and the and the client all the time and that's i think the whole edge thing is is gonna be it because then suddenly based on where our user is and their context that the edge knows about you can do personalization very easily as well absolutely but but to get to that point we also have to solve other issues probably or somewhere they're not issues but they're just how the internet works today the networks and the protocols we have at our disposal um when we have a broader adoption and availability of certain protocols for everyone or for them the most part of people not everyone will have access to internet ever unfortunately but um yeah when we have the largest population uh having access to a certain types of protocols then it's going to be easier to to standardize this type of architecture yeah exactly so that's my my take on thanks and this is when back in the back in the day i um started speaking about usher static web apps and this is an interesting thing that i am going to take the moment to you know do a plug azure has these static web apps that is a jump stack dedicated service right we know that not long ago heroku decided to [Music] be done with a free tier and so you cannot deploy applications to a free tier anymore and when people were recommending other services where uh developers could migrate their workloads they never mention usher actually it's actually it has a free tier for static web apps and it's fully integrated with github so there is key tops and and you have actions and well what i noticed that people tend to forget about azure static web apps is that it's integrated very closely with azure as in if you need authentication or something else that there's a service that exists in azure you'll be able to just get it because if you now go to another system and do this you will not be getting all these things and so somehow that marketing failed i don't know what happened there i'm not sure if it's failed at all like how much is this used it is broadly used it is probably that's just me then um i agree with you that we probably need to raise awareness um but it is very very used in the in certain um enterprise context i think that where we because they know about it it's part of the usher offer and if they are using other services obviously they want to integrate and have the full back but um it's true that we need to make uh developers more aware of the of the of these amazing capabilities that we have with us and this is what i want it might be a slight hump to go on to because azure sounds like this enterprise amazing thing that has everything which it does it's really cool but maybe that alone for beginning developers might be a bit of a big step i can kind of imagine it but i'm i'm sure you're going to show us some cli action that is very simple to do and then suddenly there's no hump anymore absolutely so i wanted to show this i did this live demo at the at this conference as well and but we also have an amazing vs code extension that allows you to create and to build and to deploy or to provision and deploy a static web without ever leaving vs code so we know that developers love vs code right and oh yeah that's that let's be clear i have not considered anything else for a long time since vs code came out absolutely so i used to work with other um editors and that i i don't even consider it anymore so exactly this is basically the architecture that we want to implement so we have the code we have the github actions and we are going to create a cicd pipeline so we can build and deploy our code and provision the the actual azure static web apps resource to azure and what happens is that um also something well has apis out of the box so if you write a a note function an ojs function in a folder called api it's going to automatically um create the resource file oh that was my next question oh nice so that's that's relatively easy just like when you use next gs and you upload it to yourself that works yeah it's it's just detecting that if there is an api there and creating the resource for you and deploying oh no look at this jiri is saying webstorm rules which it actually does but maybe not with this crowd here i mean i used to to to you it's actually great i've used many intelligent yeah i used to use intellij products i i don't like passion anybody because i did a great amount of usage of those products and and enjoyed them and they were good for me at the time but once i discovered vs code i mean the amount of extensions you have in vs code against any potential provider or service yeah or any anything you want to do linkedin i don't know whatever what i kind about something like webstorm is that it's a little bit more ide driven like it's a little bit more intelligent like you would have in visual studio basically but it's it's it's in that sense it works really really well but i kind of like the lightness of my code editor personally definitely so we have about like 10 minutes you want to maybe just show that stuff or is this more important to go over what do you think no no this is just again what happens under the hood when you are creating a static web up um and yeah and and the the misconception of not being so easy right so let's not sure that it's actually easy if you use the right tool and we have the right though because i remember giving a conference talk a few years ago and i had never used azure i had never used the serverless functions and stuff and back then it was in version 2. so it all just came out and um i had this vs i had visual studio so i had 140 bucks of money in azure back then so i started trying it and it i had this aha moment of oh the internet is actually so different than i thought it was when with my monolith hat on back in the day you could do so many cool things now oh my god i loved it it was really like it opened up all these possibilities and actually my first experience with all this was before azure web apps existed but it did have this all this azure with surface functions and stuff all right we go and we're going to go from from creating a project i did this at the at the conference like i remember okay so we're going to install a few okay cool cool so yeah we call it azure view project and let's hope let's hope i'm logged in everywhere because was something that we were discussing before the streaming yeah exactly i wasn't sure i was locked in everywhere so i'm gonna say no no no no for the sake of brevity and now it tells me like this is this the the i'm still using right npm or whatever no um it tells me okay now run this say d blah blah blah let's go i thought i was already in the in the russian cli then i realized it's too late for me let's let's install it install first yeah exactly yeah because the other time i think i did i i also got confused by this and i was like oh yeah i remember i'm already in the swazili i was like no i'm not i'm just i'm just creating the application i'm gonna have you you're so fast i'm not a new user well this is feet right yes exactly and it was not a big package i think no no no this is like the the default yeah okay so we are going to open a new tab and we're going to run sua that's all we need to run static web apps oh static web apps okay here we go it tells me um that there wasn't a configuration fine and so it runs so in it let's actually open um code so we can see it uh huh so we have the bit configuration and okay i was expecting it did at the um configuration but it's okay yeah well you haven't run in it yet right because i it's it's asking me to choose a configuration okay oh yeah yeah now you probably have it so you just say yes now i imagine yeah we haven't rendered it yet and so now it's asking me if the settings are correct before it writes and it seems like they are and by the way have you seen it's detecting the framework yeah it's very cool i love that and it's giving it's telling me like oh the this is the app location this is the output location the api location i don't have an api right right now but not yet these are the things that they were going to use also for the um actions right for the workflow so yes they are correct and this is when we're going to this is when it creates the configuration over here right okay oh yeah there it is perfect so that that's not super complicated actually this was like five things you had to do or less yeah okay so you can already deploy look at that yeah so it tells me do you want to deploy your application now let's hope i am logged in and there is no so it's running it's asking me okay what's your um tenant that's my tenant for my subscription it successfully locked me into action authentication complete whoa would you like to create a new azure static web apps project is it sim yeah so i say yeah let's go just a name because maybe i wanted to recall something else i actually could have added a um prefix and v i usually do this so then i don't have to search my resources all over oh yeah exactly exactly and now it's so what is it downloading now what is this oh it's it's downloading a binary that is called that is the static site client but this is because i had it outdated because i haven't used this for a while but if i if i had it updated that wouldn't even happen so this is going to take us gonna like people commenting like that stuff i paid what's up nice to see you so it's in progress what it's doing right now is provisioning the resource for me having to i didn't do anything i didn't go to the portal i didn't do and this is huge for me because when i go to the azure program i don't know what to do so that's cool oh it's done how about that project completely deployed to let's go and laptop you did it holy oh sorry i shouldn't be saying these things because we might have american viewers but between us europeans awesome how about it that just worked and so yeah how can you now manage this stuff do you or maybe you don't have to like if you now do an update in your code do you just run an update thing in your cli and it's done you have all these information let's go to the soi cli but ideally what you what you want to do right now is you want to push this to github right and create a workflow file over there and so deploy um if i had a these are the multiple options that we can pass like if i had for example um an api i could also start it or i could define the distribution oh okay so you can kind of choose a little bit more and you choose your notes version things like that yeah that's cool yeah so what you're asking me basically is if it's going to be watching what i do for example yes i don't know let's check let's check this part let's change the component you've successfully created a project with blah blah blah yes exactly cool okay let's see what now so now there's a change so how would i how i would do this now in my normal workflow i would just push this to git and because there's a change in git something listens on the other end and it just rebuilds right so what i would do right now is like just uh it doesn't have right oh it's already so i would just push it to to get i i don't know if i'm logged into github yeah so i'm not gonna have to do it now but so that's what you could do but what would you propose you would normally do through with actions github actions yeah absolutely so you go to you go there and then what you can do is you can even switch to the um to the github um to the github um cli if you have it you can create a configuration file from this another from this other cli and push it and then you have already ci cd set up and that's it that that's it that's going so because it knows about the swa config it will set it up for you oh that is a nice integration you know what i might have to revisit this for my youtube channel and just have a look because it's it seems this seems very smooth it's very interesting yeah i i think we are creating by the way yesterday my friend was him who uh led this this um cli from an engineering perspective he opened a pr that is going to make this faster the server actually at a 72 game 72 game yeah and so is that the professioning that becomes faster yeah that no that's that's the interaction between the cli and the deployed website okay that's awesome how can you in this day and age still make 72 percent gain that's crazy he moved everything to rust no no no no it was actually a protocol thing and it was a a connection that was getting closed and by keeping it alive that was already a big game so because it's open source we can totally yeah you can just find out yeah i've been in projects sometimes where they didn't close a connection while you were browsing the website and this was one of those monolithical systems that became headless but like mid midway headless so they didn't really know and they didn't close a connection and so when then 10 people would hit that page it would open 10 connections and keep them open and not close them and if 100 comes it would ddos itself right so connections open and closed are very important definitely but um it depends on what you're doing if you are sure if you have a multi-step being executed you probably want to keep it live for the time that this is happening right so you don't have to open three or five or ten connections exactly so natalia thank you so much for your time you're welcome and the lovely slides and talking about all this stuff and coming on here because this was the first episode so i was just i'm so happy we got like pro to help us out so thank you very much you're welcome i i just want to encourage people to play with microfrontends and compose their architectures in a different way with the user and the user experience in mind exactly and i think you conveyed that very well and you're also very impressive by saying yeah i own things like vs code and all the azure stuff like that developer experience in azure i want to to make i know to clarify that that's great thank you so much for joining us thank you and is there a last thing that you might want to plug that people should be doing things like that no but i i want to mention that a lot of our tools are open source and we are always happy to have to have contributions and of course find me in twitter talk to me if you have any usher questions related to a javascript workflow or application then i may be able to help you thank you so much and let me just remove the screen and then look at your name tag and that's your twitter exactly that's my channel my twitter my git have my my everything your everything exactly and the same goes for me here here yes all right um thank you very much and i hope to see you soon cheers thank you bye fine right yeah finding the