.Net Foundation Virtual User Group, 17 Jul 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Building Cloud Native .NET Apps on Azure

Description

In this session, Swaminathan Vetri (Swami) & Viswanatha Swamy(Swamy) will talk about building cloud native .NET applications on Azure. You will get to know how to architect, design, develop and deploy Full Stack .NET applications on Azure. You will get to learn about the below technologies

- .NET Web API
- .NET Blazor WASM based Single Page application (SPA)
- Azure App Service
- Azure Static Web Apps
- Azure Key Vault
- Azure SQL
- Azure API Management
- Redis cache
- GitHub Actions for CI/CD

This session is run by the the BDotNet user group (https://www.meetup.com/BDotNetUG/) that hosts meetups in Bengaluru, India.

A

Hey good morning and good evening, everybody thank you so much for joining in on it. Another edition of b dot, net user group meeting all right and I hope you're all doing good, safe and sound and healthy right and- and I want to start with the thank you note, because there's a lot of things going around and then you chose to log into this event and then try to learn something new. So thank you so much so before we jump into the event.

A

I would like to give a brief introduction about myself and then about the user group and then we'll get into the event all right, so myself swami. So my name is my full name is swaminathan vitri. I work as an architect at musk, I'm a microsoft mvp under developer technologies, I'm the organizer of b.net and exactly who's a group made in bangalore and that's my github handle and I do write blogs occasionally at my at swaminathinvity.in and if you want to follow me on twitter, that's my twitter handle all right and this is about be.net.

A

So video is one of the oldest user groups in bangalore way before all these meetups and the things started. uh It's been more than a decade since it's been running in bangalore and I'm fortunate to be associated with it for the past seven years or so uh and yeah. So these are the different ways you can follow us, so one is meetup.

A

We also have presents on meetup.com and then on facebook and twitter and video.net.ion is our website, which talks about the different events and the previous presentations and things like that, and we recently started youtube channel. So you can also subscribe to us and then we would upload all these videos into those channels too, and we also have a discord server available.

A

We can also log in you can also connect with us there, yeah and of course, next things about the sponsors. I would like to thank our sponsors jetbrains, who are sponsoring us with their licenses and then microsoft and dotnet foundation for giving swags and providing streaming support for this event right.

A

So I would like you to like you all, to be aware about dotnet foundation, so if, in case you're not aware of it, it's an independent, innovative and always open source to promote that dot and foundation is an organization which is created and what it fosters is and vibrant.net ecosystem. So with dortmund foundation, you can basically take help on all these aspects.

A

Let's say if you're working on some of your opponent's projects, if you want to understand the legal or the ipip stuffs and about the license agreements to set up the country good license agreements, or how do you set up copyrights, trademarks, etc? All right, if you're running any dotnet based upon this projects, you could take help from the dotnet foundation and then get these things done and then about even about the time. What what are the kind of focuses practices you want to put into your application?

A

You could also not the applications of the project you can also. You know, help take help on those lines, be it setting up the governance or seeking the feedback from your customers or providing any kind of mentorship coordination, etc. Right apart from this, if you're, if you're running meetup groups and then if you want to if you're running different events, dotnet specific events etc, you can also get the help in promoting and then getting your event marketed, etc, and not just that you can also get some amount of help in the hosting part of it.

A

If you're looking for some space to run your ci cd pipelines, you can basically get a devops support as well right and yeah. That's the github, url github.comslash.net foundation and dotnetfoundation.org is the website right and how to get involved. So this is one slider which I like it. So there are different ways you can get involved into involved with the dotnet foundation.

A

So if you want to become a member member, just go go to that link and if you want to attend other meetups, which are hosted by dot and foundation, that's also possible or if you're running a meetup which you want to be, you want to make it to be part of dotnet foundation.

A

You could also do that and or if you want to get the newsletters you can sign up for that and, like I said, if you are running a project, if you want to contribute your project, you can also go and look at contributing to the existing projects or, if you want to nominate your project also, you can reach out to them and then go through the process and then get your punches project nominated and and made it part of under the dotnet foundation umbrella, right and next, the speakers so we're going to have two speakers today.

A

One is me swami nathan, very another is wishimata swami, so we both will be talking on there on the cloud native building cloud native applications on azure and it's a pretty sleek agenda. Today we don't have multiple sessions, so this is even work through which I'm doing currently and then we'll be starting the session by 10 5 a.m. And then it's going to be a comp, a single session. We are.

A

We are planning, like a more of a court jam kind of a session, so initially 15, 20 minutes I'll, be having a bit of a presentation to walk through about the context and on what is cloud native app and things like that, and then there will be fluff demos for the next couple of hours and we'll have q and dedicated in the last 15 minutes. And finally, we will also need to take some feedback, and then we have a raffle and yeah we'll do the wrap up by 4 30 p.m, ist.

A

So this this is a bit of even a ticket, so q and d will be handled at the end of the session. uh So but that doesn't mean you'll have to be uh resisting your questions.

A

So keep posting your questions on the youtube and if you find it relevant to the topic what we are discussing, maybe we could pick a couple of them in between and then we could discuss about it because today is going to be like an um like single session which is spread across or which is spanning over last for the two hours right, and we would like to hear about your experiences so please tweet about it using the hashtag, b.net and yeah, there's also a raffle. So I want to like.

A

I wanted you to see the feedback first and then whoever has produced feedback. We will do a raffle from the people who have submitted the feedback, and then we will be giving giving away three jet branch licenses right. So with that yeah all right, we could get started in a couple of minutes. I think we are already on time, so maybe I will quickly get started with the the presentation.

A

So yes, the topic for today is building cloud native apps on azure, cloudnative.net apps on azure right and it's going to be a joint session. Like I said, uh one is my myself who is for me and then other is maybe let me quickly bring bring him to a stream, and then we could, you know with starter.

B

Good morning, how are you good morning.

A

All right, so uh I'm having a bit of network and the power issues today, so I couldn't turn on my videos, I'm just not just going to be on audio today, all right, okay! So let's get started.

A

Sorry about that, for.

C

A moment I was thinking that we have three speakers today.

D

Yeah, I thought what happened.

C

A

The work from home setup right, I.

C

Know you, your daughter and me looks like we are three speakers today.

A

Right, yeah, okay, so I was told that the power is going to be coming only in the evening, but anyways I'll run the presentation and then we I'll hand it over to vishwa who can run the demos after some time right all right. So this this is a big brief intro about myself and on him. So I just talked about myself. So I'll just try to talk about only vishwa now, but he's also called swami.

A

So he's uh he's working as an associate architect at ais and he's a c-sharp corner mvp and an mct microsoft, certified trainer and he's an organizer of.net learner's house, that's under meetup group from hyderabad and yeah he's also a blogger and a speaker, and if you want to follow him on twitter and github, that's his handle right. Okay, so the agenda of today is: uh is this right? So first we'll start with understanding of what are cloud native apps and then we'll see. Why do we need to build cloud native apps right?

A

Let's understand the: why parts of it and then let's also see how we can build the cloud native apps right and then we'll also look at what are the future azure services which we can use to build cloud native apps, and then it's going to be a lot of fluff demos today right. So what is the cloud native app right? If you ask anyone in your organization or or your friends also, you will get different flavors of definitions for cloud native apps right.

A

So this is one of the official definition from the cncf, which is the cloud native computing foundation, so cloud native technologies empower organizations to build and run scalable applications in modern dynamic environments such as public, private and hybrid clouds right. So the some of the points which I've highlighted here are the ones which which are more, which are more around cloud native, specific things right. uh That is what we'll also now understand over the course of this entire session right, so containers service meshes microservices and immutable infrastructure and declarative ap.

A

These are some other concepts which will exemplify this approach, and we will understand how we would be applying some of these right, if not all of them. In this session, and today we will be focusing more on the on on especially azure pass offerings, and then we would discuss about the containers in other other aspects in the in the later sessions right and on these cloud native apps. What do they do? They enable loosely coupled systems that are resilient, manageable and observable all right, so we will.

A

We will expand on these different terminologies over the course of it, and and when we combine these with robust automation, I think we'll be able to make high impact changes frequently and with and with high predictability right. That is what would make a particular application, a cloud native app right and why building it? I think we just talked based upon a few of things. One is autonomous subsystems right, so cloud native apps are composed of autonomous subsystems.

A

What I mean by that is an app could be built with multiple other different subsystems like it could be broken down into multiple pieces, and each of those pieces can be built, developed deployed, tested, tested and deployed independently, and that's what is an autonomous subsystem means right, and it also brings in agility right, so you will be able to move faster and that's with agility. You'll also get a faster time to market right, and since you are trying to break a system into multiple subsystems, you get an option to scale them independently.

A

Also right so achieving a hyper scale is, it would be, would be one of the one of the greatest. uh You know advantage of building a cloud native application and, of course the performance also you'll be able to make make your make your application highly performant, and you would be able to choose different parts of trouble. Now you could be able to choose different performance performance benchmarks, are the performance uh or to say performance numbers for different parts of your application right?

A

Okay, so now how to build cloud native apps right, we just talked about what is the qualitative app and then why? So, let's talk about how we can build it right. So there are some certain principles which are built and it's called a 12-factor app right, we're going to talk about what are the different 12 factors and then how they would be helping us right. So in a 12-factor app.

A

The first thing is the code base right, so that has to be only one code base which would be tracked into a version control system, be it gate or subversion or whatever be right, but in the current world most most likely uh at least we would be using some flavor of a git version control system right.

A

You should have your code put in the version control system, and then there could be multiple deploys happening when we say multiple deployments, it could be with deployment to your different environments, right your development, environment, testing, environment production or pre-product production, etc. Right all of this has to go from a common code base right and then the dependencies, the second part, is round. So how do you handle the dependencies right so we'll have to explicitly declare and then isolate the dependencies right?

A

So what I mean by that is: let's say if your particular application, let's say, for example, if you, if you're, building an application which is consisting of multiple subsystems right, as we saw in the previous slide so and there could be certain things which are common across these subsystems right. So those things also needs to be called separately. They need to be isolated and then they need to be put in separately.

A

That's one part, and the second part is around: how do we, let's say: if building a dotnet application, we would need a dotnet runtime, which is one of the dependencies right, so it has to be packaged as part of your application. We should not. There should not be any implicit uh uh expectation of independence. Let's say if you are going to deploy your dotnet application in a server.

A

We should not implicitly understand that. Yes, the server would have an uh dotnet content running in it, and then we go and just deploy it. That's how we traditionally used to do all right. What would happen is yeah. The development people would would kind of uh create their own, uh create their application in their local environment and then they're not bothered about what is going to be present in the actual actual production or development environments. And then, when you deploy it, you find issues right, that's what we would want to avoid.

A

So that means you need to declare all your application dependencies and you are in running execution, dependencies, etcetera and all of them needs to be packaged together, or at least they need to be declared declaratively specified here right and third parties about the configuration. So how are you going to manage the configurations in the environment? I'm sorry. I think something is going wrong with my animations, so it keeps repeating uh so configurations.

A

We don't want to hard code any of the configurations into the into the code base right. So whatever you think would be specific to environments that could change dynamically. They all need to be maintained specifically in a configuration files. So one of the typical example is, let's say: if you're building and micro service or an api, if it's dependent on multiple other apis, you would need to store the urls of those apis right.

A

One easier way to put it is just keep all the urls in in the checked in into your are added into the constant files are so. But that means every time you want to deploy into different environment. You would need to modify your code and then you have to deploy it, which is a violation of 12 factor, app principle right.

A

So what we could do is kind of externalize it and then make it in the configuration, and then these config should change dynamically as part of your deployments and then to deploy the same code with multiple. You know, configurations in it and the next one is backend services, so trade backing services attached resources backing services are nothing but the dependencies. What what we have, what an application depends on right. So one example of a dependency could be: uh let's see if you're having a database as a data database as a dependency right.

A

So that also needs to be handled separately, so these databases uh are not part of. Basically, if you want to process your your, you have to use any kind of persistence, you would be using a database for sure and those databases are treated as an additional resource and they're, not part of your existing one, and it's not just about database. It could be like you talked about the the common component or shared library.

A

What you want to use across different components right, they also would be treated as a separate resources right and then bill run, build release and run. So we need to be able to differentiate the build stage, release stage and run stage, so build stage is very for all. We have to relate it to our continuous integration in one month. Right, build stage is where you would do your compilation and then running your unit test, and things like that. So you try to do a compile and then get the get the package out of your application.

A

If it's a dotnet application, you'll be creating and maybe a dlls and then create an msi or whatever, and we used to do earlier with containers, maybe you would create and docker image or so right, that's the build stage, and then, once you have your build a stage and your release stages should be separated out.

A

So that means whatever the coachings you have you keep on building it and whenever you want to release it, you should have a different release stages and from that early stage your app should be released to the appropriate environments right and they when they are released, they and the run stage, is nothing but the execution context of that right. So wherever it is going to be running, that is where that's. What is the run state and all of these needs to be separated out.

A

Excuse me, and the next part is about the processes. So the process is uh basically we need to execute the app as one or more stateless processes right. So that means any kind of sticky sessions and then maintaining the states is is a violation of at all factor principle right. So because uh the application could uh application should not maintain state, that's what a stateless obviously means right, so every request could be processed by different different version or different instance software applications.

A

So you should not be maintaining any kind of and sticky session or any kind of stickiness with your with the processes. Right, that's what is about the processes and moving on. There are also other six things which is on port binding. That means, basically, your app should be listening on some of the existing ports. This is more applicable for your web.

A

Apps are so right where you would expose it on a port and and then your application would have an application server, maybe like a web server or so that would run on the port, and then you were all your calls would be coming from that port. So with this, what would happen is let's say if you are porting your let's say, for example, dot application, default.net app runs on localhost 5000 port right. So what would happen?

A

Is you were having exposing and dotnet uh whatsoever on block 5000 and tomorrow, when you are trying to put this application into a different different area, different instance or different environment? And if you want to change, that's the port and then you should be able to still access your application right. So that's what we're getting decouple these things now, where the application is exposed versus how it is running under food is kind of you know.

A

Decoupled here, part from the concurrency, uh so concurrency is how we can scale out the individual processes right, for example, there could be multiple types of workloads which we could create like something would be like in web processes, which would take the request, response kind of requests, and there could be another thing about the worker processes which would be doing some sort of background processing, etc right, so we should be having the flexibility to make these things scale out independently.

A

If, at all you see there is a lot of requests coming in for your http based workload, you should be able to scale them out separately right or, if, if you want to do a background processing in a more in a compute heavy environment or sorry, you should be able to do that. Also, that's what concurrency talks about right and and the next one is around disposability disposability means. Basically, we need to maximize the robustness with the fast startup and graceful shadow.

A

What that means is because these apps could be uh are not going to be definitely running in the same execution, environment right, they could be deployed into different different or they could be moved around. For example, if you're running on the cloud you don't know in which machine it is going to run, right and and azure would take care of moving this into from one of your on one of the servers into different servers for any kind of reasons or so right.

A

In such cases, we need to ensure our application is having a much faster startup and also it should if there is any any any shutdown happening for your application for any any underlying operating system related issues or whatever your application should be able to gracefully shut it down. It should not have any kind of again stickiness with the with the running environment.

A

Right, that's what disposability is talking about. I know these are high level concepts, so each one we can talk about uh for for us together to get a better understanding, but I'm just trying to you know, explain you and give a high level overview on what what are the different different principles. We need to keep in mind when we are building a 12-factor, app or, and how that perfect app would help us build a cloud-native application.

A

All right, an expert is around the dev or broad parity. So, basically, we need to ensure the development staging and broad environment. Everything should be similar as possible right. This is this is one way we could reduce the issues like hey. It worked in dev environment, but when you put it into the production or when you put in the qa, it failed right that could be if the difference the environments are different. We can't really predict what are the? What are the reasons of the failures right? You would definitely get an non-predictable application.

A

Basically, predictability will go down right so to ensure you are able to identify the right issues in the right environments. We need to make sure the dev and the product, and everything is, is not parity right and then and and the last next one is around the logs part. So traditionally we all used to write the logs in a log file in the servers right. That would work only if you know only if your execution environment is not going to be changing.

A

Let's say in an on-prem kind of setup where you will have a physical server or a vm where you go and deploy your application in iis, and then you will write the logs into that particular server's file system right, but in the cloud native world, it's not going to happen. That way. Right, like I said you were, your application could be moved across different instances of vms or different servers or different data centers for that matter too.

A

Right in such cases, if you're, if you're, writing anything into the file system, it's going to be very difficult for us to uh to look at the logs and then understand the reason what's happening or doing any kind of troubleshooting right. So what we need to do is you need to start considering logs as even streams.

A

That means you should your application should have the ability to write the logs into the into the standard output or console, and then we could have a common process running in the execution environment that could extract these logs and then emit it into your uh whatever the log block processing system, you use the splunk or el car, whatever stacks right, the idios app should not be writing anything into the file system. It should start emitting logs as different streams.

A

So as and when the app execution happens, your log should keep emitting and then there could be another process. It could read the logs and then it would push it into the into the appropriate logging system or the monitoring system right, and the last part is around the admin processes. Of course, we, though we try to do any kind of automation, everything there would be some one-off processes or scripts.

A

We would need to run right and those needs to be separated it out, and they also need to be run or in the same environment as how your application is or where your application is running and that will again ensure the parity and then we can. We can avoid a lot of uh you know, discrepancies issues due to enrollment differences, etc.

A

Right excuse me, so that's that's about the 12 factor. App and next is now uh moving on right. So one of the ways to build these cloud native app and following the 12 factor app is the cloud cloud is the uh cloud. Is the microsoft is following by following the microservices architecture right, so what are some other characters of micro services?

A

So some of these we already talked about so each microservice would implement a specific business capability within a large domain context it in the in the case of an e-commerce portal, or so you would have multiple kind of functionalities like search and then could be. Something could be like a cart and something would be like an order process etc.

A

So each of them could be split into different microservices and then they all of them together will form an application right and- and one of the main things is about each of these can be developed autonomously and can be deployed independently. Also right. That brings a lot of flexibility right. You can reduce a lot of friction between different teams and then each team can build and on test and then deploy and release on their own on their own release cycles.

A

If at all, you are doing in a monolithic fashion, you would need to have a proper coordination between all of them right and also each each is a self-contained encapsulating, its own data storage. Basically, each microservice should own its own data and then the programming platform.

A

uh It could be whether it's equal or no sql and and and when you talk about micro series and microsoft- should have its own bounded context identified and it should have its own data store, also identified right and the next one is around each transition processes and communicates with others using standard communication protocols right so now that we are breaking a single system, our single application to multiple microservices, one of the important things is around the communication protocols.

A

Right it could be an http based protocol, it could have been even driven or it could be message driven or it could be web sockets right. But the important thing to consider is the communication. That is how all these can can work together and, of course, all of these different micro services would compose, uh uh but basically all work together and they compose a proper application.

A

And this is just to give a brief about how the deployment would be different between monolith and micro services, right so with the monolithic application. What we do, it will be typically an entire kind of an application. It be used to have your business logically or a front-end, and then your back-end, and then your database and things like that right.

A

So on what we used to do is we used to deploy the entire unit as a single unit, and then we put it in one vm or so, and if you want to scale the application, we would add additional vms and then do the entire module deployment there and then put a load balancer and that's how the scaling would work right. What it brings is, every time you want to make a change.

A

You will be deploying the entire application, for example, if a search model is getting changed, even then you need to deploy the entire application where that might not be any changes related to other modules right. So that brings its own challenges so with microservices. What we would do is we can. We would have the flexibility to create different micro services. Basically, we break a single monitor into multiple micro services and we can deploy them independently and we can scale them also independently.

A

So that ways, let's say if, in a typical ecommerce lot of people who do a search site, but not everyone, would go and buy the same item also, so obviously your search should have and much better scale available, but on the other side, when you do an order crossing also because these are kind of background processing systems, you could have a different scale maintained for that all this would be possible if, at all, we are following the microservices approach.

A

If not like I saw in the left-hand side, you would be adding more vms and then you'd be spending more on the infrastructure and then you'll be having a lot of issues with respect to your deployments, etc too, right and now so talking. I think we understood what are cloud native apps and then what are the different uh uh total factor, authentication factor, app uh properties, etc. Right now, whenever, whenever you're building a microsoft, what are the design constructions? We need to have that's what the here we are going to talk about.

A

It's one part is communication, so communication, like I said, is, is one of the key, because we need to ensure how or who can talk to these. What are the different microstructures that can communicate with each other right and how are those communication going to happen? Is it going to be by an http? Is it going to be a rest, or is it going to be a grpc, or are we going to use an even driven system may be putting as a kafka or or even hub, or anything for that matter right.

A

So this is this. Communication plays a key role when you're architecting your cloud native app and because you don't want to make it too chatty. At the same time, you don't want to violate the principles of micro services right telling you you don't want to directly go and access anybody else, any other bounded contact database also, so we have to clear cut. We have to identify clearly. How are we going to enable the communication between these different micro services and resiliency right? That's another important aspect, so the resiliency uh means what has been basically.

A

Instead, we we our application, like I said it's going to be running on the cloud and then could be moved around in different different execution environments right. So if one of your dependent dependent micro services down also, how are you going to handle those those kind of things, and how do you make it more fault? Tolerant? How do you make it more? uh You know resilient to the to the dependence to the to the issues with the respiratory, different systems, etc.

A

That's one of the important pieces and then the third part is from the distributed data. So here, since uh the data is also going to be owned by indus micro services, that would be challenges in how we would be querying the data and then how we would be maintaining transactions with respect to the data et etc. That's another thing which we need to mainly focus on then consider and also the identity.

A

So we also need to understand what are the different micro services and then what is the induced identity and then who can call who other so, maybe which microservice can call which other micro microservice right? So we can't simply open and open up all these. Let's say if you have a bunch of microservices, maybe 10 to 12 in your application. We need to know. Okay, maybe so vca should be able to talk to only b c and d, so you say should not be talking to maybe e f and h right.

A

So we need to know what are the identities and then who can invoke, uh uh which is the microservice right and, of course, security is, is the base of all of these things we need to ensure all of these are happening in secured manner. You don't want to have any kind of an non-https request or anything floating around. You need to ensure there's a mutual trust, enabled and there's a zero trust network built, etc right.

A

This is also another one of the important considerations what we need to have, while, while working with microservices architecture and and backing all of these would be like your devops processes. Right devops is not just about running cac, it's also about uh provisioning, your infrastructure. How would you make your infrastructure as a code and make sure you're able to get immutable infrastructures right? If there is any sort of an issue, you should be able to bring up your own infrastructure as soon as possible. That would be possible only with infrastructure code.

A

It maintains some sort of your infrastructure in code in either arm templates in azure or terraform or polymer. There are different providers available to achieve the infrastructure as a code, and that's one part second part is also around. How are we going to build release their application?

A

So that's where ci and cd comes into picture, we will have to have a convenience integration build so that, as and when developer commits the code, we need to get the faster feedback, and then we also need to have means to deploy on the release your application on demand any time. It's all. These and also about the testing process, test automation, bringing in uh functional automation, on eap, automation, ui, automation and things like that right all of these would form under devops.

A

So when you are working on or whenever anyone is working on a microservices architecture, you need to ensure at least these are the different things which needs to be considered. This is, of course, not the exhaustive list, but these are the bad minimum things. What we would need to consider for sure right now. uh I think we talked about a lot of the rate now to understand what are the different azure services that could help us in building cloud native apps right and before I jump into services. There are.

A

There are multiple ways you can build a cloud native app right like a test upon the first slide right. One part is, you can use the can completely converge as your pass offerings like using app services and things like that, uh and the second part is around making it serverless. When you go to serverless model, you could use azure functions and there are other host of serverless offerings available and the third way to do it is using containers.

A

If you want to use containers, you can use docker based containers, and then you can use an azure container instance or kubernetes service, etc, etc. Right so first thing is azure app service. That's one of the easiest way to you know, get us onto the cloud. So, instead of we maintaining our own, it's basically azure tasks, offering it's a platform as a service. So you don't you don't need you don't have to worry about the underlying infrastructure. All you need to worry is you would need to define, is what is the underlying runtime?

A

You are applying to have your application running on user.net5.net6.net3. Whatever right, you need to just specify that and then app service will take care of it. App service will run the execution for it and then uh another important thing you might need is an azure api management. Let's say you've built a bunch of apis like web apis right and and how do you manage those apis? How would you ensure that routing is happening appropriately? How would you ensure we are, we are able to?

A

You know, uh uh perform authentication and authorization, and and uh and things like that right, that's where ap management would come into picture and now that if you have built multiple apis and then, if you want to make it like an uh jio replicated, what I mean by that is, let's say: if your customers are across the globe there, you have a global customer base and if you want to manage your applications and the data closer to your customer location, so, for example, I'm building an application which is for which my customers could be accessing the app from either from india or from europe or from the us right.

A

So these three are primarily three different regions, so maybe for the india I could have have my micro service running in my india data centers and, if I for, for the europe customers, maybe that the the application calls should go into the europe for specific endpoints right and then for the u.s. Maybe it could go into their specific end points right.

A

So if you want to do all of these things, you can use something like azure traffic manager and you would put a traffic manager on the top and then that would have the jio specific routing enabled there.

A

So, whenever your us customers hit your api, it will route the calls to the your specific micro services and, and that way you could reduce latencies and then you can handle a lot of things right and- and these are some other things which we could use and again I'm talking about the uh containers in kubernetes, which is a vast topic by itself. You could also use azure kubernetes service, it's a managed, kubernetes service that could help in uh orchestrating your containers and yeah.

A

If you're not looking at a full-blown kubernetes service, you could very well use azure container instance. So these are like standalone instances and then they would also uh they would. They would also be able to help you running in the running docker containers of your applications and yeah. Like I said for serverless, you can use azure functions, so that means you will you these?

A

The functions are primarily like very, very my new entities right, so you, if you want to do one specific operation, you put an azure function in it and then you will run it up and the functions have multiple multiple triggers and multiple bindings, so triggers are the ones which would trigger the function on what basis the function should run right. It's not like a typical web application.

A

Where, let's say, if you are running an on-prem, you will be deploying application iis and then it will be always up and running and whenever a request comes in, it serves a request and then it will. The process would still be there right, barring the is timeouts etc, but with azure functions it will wake up whenever there is a need and then it just goes down. So that way you do. You can save a lot of money in in the execution cycles, etc. You'll.

E

A

Paid for whatever execution cycle you are, you are using there right and, of course, to ensure we are accessing things in a secured fashion. We would also we can also use key vault and yeah. Cosmos db is one of the examples of a back-end store. There are also multiple offerings. You can also use azure, sql, azure postgres azure for cassandra. Is there azure for mariadb mysql, a lot of different data store options out there so and yeah? There are many more applications of them.

A

Sorry, many more services which you could use from azure to build cloud native apps right, like I said, the focus of this session is more around the past offering so we wouldn't be touching on the kubernetes or the other functions or anything else right. So now, uh what are we going to do through the end of the session right?

A

So we try to build some of the use cases, and then we are going to have we're going to explain how we we are trying to solve it by with a cloud native uh kind of an architecture here right. So it's a very simple use case. It's nothing fancy! So the business requirements are like this: we should have enriched interactive client applications for the end users, which is what everyone needs now right.

A

That means you could technically, you would need an sba or so single page application, and then the client set app should be able to load the data dynamically and then on demand. So that means it's not a static application, so it needs to be wired up to some of the backend apis which could serve the data right and yeah. The data should be persistent transaction level and the ability to query and the pull reports, etc. So this is the database right, so it should.

A

We should be able to basically perform some credit operations into your database and then we should be able to query and then pull the reports from that database and, of course we need to enforce security. So we want to have some sort of role based authorizations for the users and- and then you see, these are some of the high-level business requirements right, but we don't just build and for focus only on these.

A

Typically, if you're in enterprise, you don't hear like the nfrs and other things, which is a non-functional requirement, that's what I call it as a technical requirements here. So another thing is security aspect we need to ensure we are securely handling the application configurations on the secrets right and the second part is we need to ensure our aps are able to give the responses in sub. Second, all right and yeah. We need to also have an end to an observability of the entire solution.

A

So what I mean by observability is you need to know how the traffic is going between different micro services or the different parts of your application from your ui to the back end, and then, if they're back, an ap is calling multiple other apis. How is that call happening all those need to be able to monitor and then also the logs of it into and tracing? Also, we need to be able to do so that the moment there is an issue we should be able to go and troubleshoot, what's really happening there and yeah.

A

Of course, we would want to have an end to an automated deployment of the application, which is also one of the important principles or the design which would need to follow when we deal with the cloud native apps all right. So with this with these different requirements, we kind of you know we're going to look at and sample architecture which will look like this. So basically, we are going to use primary.net as our programming language so or the framework and how we are going to build this cloud native app right.

A

So whatever I talked until now is is kind of common principles on the under design patterns or the design methodologies right so which could be applicable for any kind of any kind of an programming language, be the tool factor app or the microservices architecture, or the cloud native principles, etc. Right and and even the azure services also, they could host even azure apps can host any sort of and any sort of programming language if you're having a microservice built in java or node.js.

A

You could technically put into the into the azure app service too, but the focus here is more on the dotnet part of it. So we are going to use dotnet to see how we can build the end-to-end full stack, application.net right. That means the back-end web apes are built on dot-net based apis, and then the front-end also is built on the blazer. There's a wasm, that's nothing with the blazer web assembly and both of these are deployed into into the azure right.

A

So the numbers mentioned here number two is the azure app service right and and for the security to enable the role based authorization we kind of integrated with r0, but it could be any any kind of an authentication, any kind of identity provider right. So you, if you're, if you're on in an enterprise where you as your a it, is enabled there right.

A

You could technically connect your blazer up to that and then you can get your users authorized from there or if you want to enable third-party integrations, also with github or linkedin or whatever it. You could technically be do that too, and the idea is we want to. We want the users to log in and then we would be using a jot token as a means to you know, do their authentication authorization and then the apas would honor those requests right. So that's where you see the uh api management here mock bus number four.

A

So that's going to be the entry point for for all of our consumers or the customer or consumers uh in this simple architect, there's only one consumer which is basically azim app, but you can technically think of multiple other consumers too right. You can also build up and mobile app that could also talk to the apa via this api management right and the actual apis are hosted in or would be hosted in the azure app service.

A

Here again, it will be under instance, of your app service, and the key vault is what we are going to use to secure the connection strings and everything right, because, as you can see here, the web api is going to talk to the azure sql server and also redis cache right. Obviously, we can't hard code.

A

The hardware the connection strings there, so we need to be able to put it in a secure place and that's where keyword comes into picture right and here, azure sql is the is the backend which would be used and just to enable the caching like a cache, has a pattern. We are we're trying to bring in the azure. It is cache all right.

A

So that's about the architecture. So what we would be doing in the next couple of hours, or maybe one two and a half hours is, we will start looking at how we can build some of these components and then start deploying them onto the azure from visual studio itself right and yeah. Before we get there one last thing about to talk about the cac departure you want to set up a ci cdn if you've not done it before. This is a simple steps which we could use so anytime.

A

A developer comments code into a git github I've just taken github as an example. Here it could be any kind of a git version bit bucket or get labs or whatever right. You would on a ci where you would you do when you say bill, it's just a compilation aspect.

A

You need to pull your and you get dependencies and then build your application and then run your unit test and then run your sonar, cube analysis or whatever static code analysis uh to identify any code smells then sac is nothing but the static analysis for security testing, uh so you can use polaris or any or any other such tools to understand your and are there any uh security vulnerabilities into the code etc?

A

Right, you can do all of these things as part of your ci and then once everything is ready, uh you could have another cd process setup that would try to deploy your application into the app service right. Can I try to put it in a generic way as deploy, because there are different way you could do the deployment and this deployment would vary for what are the destinations right? If it's an app service, it could be. Maybe your entire package, you could just deploy there, but for all you are.

A

You are deploying into an azure and cuban service. You will need to push it into an azure container registry and then you'll have to deploy the image. uh If it's a serverless, it could be again a different me right. So that's why you want to keep it generic, so this also would be to you. Can you know you could be relating it right? So the build part is what is the ci, which is one of the things with build and then our release and then the run part is all isolated here.

A

It's one of the principles of 12 factor app right so that that's how that's what how or that's how we wanted to run through this entire session. So basically we'll be building different components, and then you should be able to relate on how these things fit into the overall collaborative architecture and then how these things are. How are we implementing some of these 12 factor app principles?

A

Yeah? So with that quick intro, I would let which one start with the with the demos and the other things, and we would like to make it a little bit more interactive between both of us. And if you have any questions, please feel free to post in the comments and then I will be picking up from there and then we could answer over the course of the session. Also yeah over to you.

C

Hey uh thanks swami. It was uh pretty extensive interaction. So let me share my screen.

C

I'm sharing my.

E

C

D

C

Screen visible, wow, okay, so the code for this entire session will be under this vishy pylor blazervasm, apim, api sequel, readies and swami has already gone through every bit and piece so just like. I just wanted to give you this path, I'll, send it to swarming. He will.

C

Step by step to configure first we'll discuss on this sql azure and then we'll start with this web api, then integrate with azure key vault and introduce reduce cache and then move to the ap management and then bring this blaze of asm and integrate with oauth and then finally, we'll see the application.

F

C

So I I just wanted to create some content with exceptions and other things. So this how it looks I just will take quick uh five minutes uh run through on the existing infrastructure and we have another resource group which we will be using for the live demo. Just in case. You know that demos will not work on time, so we already created one and tested everything and kept it ready. So we can fall back on this.

C

So this is uh the application map where you can trace in the future episode. When we are doing the architectural thing, we will introduce the client, uh the client side app too, so that you will get a full end-to-end picture. So at this point of time we are tracking from the a beam to uh api, then sql and auth 0., okay, and also you can come and see the failure section, I'm just giving a quick, we'll come back and detailedly discuss about this just quickly.

C

So the types of failures this if this numbers are more so we can. One of the thing we can take out of that is someone is trying to hack our system. He doesn't have the token, but he is trying to enter into this one okay. Similarly, the subscription key. So these are the two red flags really red flags. We need to focus on, and apart from that, we can have this live matrix, where we can keep a watch on that. I will go ahead and use this for sending few few calls.

C

So I can see that we'll see that live matrix and let me stop sending the token so we'll start seeing the red line bumping up, see it started bumping up.

C

So if someone is monitoring the live matrix or we have alerts on this one and the moment someone tries to enter or not and do an unauthorized calls that gives us a chance to look into that red flags. Why someone is trying to do this one, and we will go in depth of this one. Okay- and I do want to show one simple piece as this one.

C

When I come into this, our radius cache- and uh this is our system and I go and do a first hit so the first it is going and it's bringing it took, 409 milliseconds.

C

So now you can come back here and do a quick keys and check that at this point of time we have all the records stored in the release and now this time, when someone tries to go fetch the data it brings in because that was the first hit, let's go and get the uh second it, so it see that uh it's coming to 339 milliseconds, because it's it's it's completely uh passing through the apm then goes to sql then goes to api and then goes to this one.

C

So as and when we do this, it takes lesser and lesser time we can check directly from the api uh we are hitting it. So it took longer. It took 361 milliseconds, so the time varies seat. It's uh it's coming down. So that's one thing we are introducing. So let me come back to start discussing on the piece by piece. First, let's go to the first piece we are going to chew. Is this one sql azure?

C

So let me walk through all those things which we have swami just give me a shout. If you are unable to hear me because uh I see um because of this rain, sometimes the internet is fluctuating uh it's coming. Actually it will be 50, but for some reason it is coming down just to give me a shout.

C

Thanks, so this is a production resource group which we have created so where we want to do uh live uh demo, and we know that to spin up this apm instance it stick. It takes a decent time of 45 minutes to one hour, so we just created but nothing in there. uh If, if we come and see this, there will not be anything here, let it let it come, and also the sequel, and I would like to flash one of the.

E

C

Use this page to define any of your resource group names, never ever define my storage, his database, uh their resource group. No, never don't use that. My and those things always please follow. This is an excellent page. I really like it.

C

They have given an all the abbreviations which we can use it, and also there is an uh corresponding or the supporting document to this is how do you choose the name? That is also the naming convention also they have given- and please use these two like you- have this uh pip sharepoint like a resource type workload, environment region and instance uh like if you see that we have also followed the combination of both like environment. So, if you come back to here, you can see that rg.

C

So that's coming from the recommended recommended uh abbreviations and we use the b.net cloud native dev001, and here we used prod001 yeah. Here we used plot001 just to differentiate. So please keep this resources very handy, I'll, collect all this in a notepad and end of the session I'll give it to swaminathan so that you know we will put it in the youtube channel or wherever it is appropriate.

C

So please use this naming convention and that gives if anyone sees that and also tagging as we didn't get time, we didn't do it, but if you see that at a high level- even if you don't recognize these icons, what you can do is like apm api management. This is app service, app service. This is application insights keyword. There is log workspace release, there's a sql server, there's a sql database, so we have pretty much clear cut by seeing the name itself. We'll know that, okay, what is that it is happening over here?

C

Okay, then, so, let's quickly go and see what is there in the sql server, and here this is the sql server. We spun up a very small resource one, but you can use uh whatever is applicable for your production.

C

Scenarios but uh video diligent to do a performance testing or a feasibility study before choosing the tier and.

E

C

And also we use the single database because we just needed a basic basic tier. We just did that and used that one and another thing is: you can use the elastic pull uh where you can. If you have set of databases, you can put all the databases in an equal but before that ensure that they all are not spiking up at the same time.

C

So if you will save some money when you use the elastic pool and and ensure that the usage pattern is not at the same time, they are not hitting the rooftop at the same time. So let it spike at different instances like I have four databases. Every database is individual databases each of hundred dtu's and if I want to put all those things in 300, d2s or 250 or 400 videos, whatever it is, those dt will deduce.

C

If I decide I need to ensure that the database a b c and d, they are not hitting the rooftop on the same time. Let them hit it on the different time and I can reduce the number of dtus in the elastic pool to 250 or 300, but that should be a very, very careful decision. If not, we will have a sleepless nights for sure. Okay, coming back to this uh sequel, this is sql server database, we'll go step by step.

C

So at this point of time we don't have anything we can I'll go into this query. Query.

C

Editor looks like my uh net. Is little slow see this okay, so.

C

Come in this uh query editor, this is a new feature.

C

Okay, so this is one of the first first security. The thing is that we are not allowing anyone in this world to touch our database, so we will come to this show firewall settings, and I just wanted to speak about a couple of points here from a security perspective, looks like my net is still slow.

B

C

God so you can see this, the minimum tls version is uh 1.2, which is the latest one. So we got additional security over there and and see this here, allow azure services and resources to access this server. So at this point of time initially it is no. That means, even if I go, and even if I go and connection string to any of the web api, they will not be able to come and hit that because we didn't allow anyone to come and hit and the client ip address.

C

We are not giving to anyone the client ip address so whether to have a dedicated private network and give access from that network. But at this this is a demo, I'm just giving an security exception. So how do you know that? What is your ip? So you can go to browser and check what is my ip address and then enter, and then you will get the same ipad address. That's for your router, which is at your home or okay. So that's 2, 30 37! So that's one, two, three! Thirty seven!

C

I can add this claim ipfs, but you guys never ever do and it's a must or needed, and another important thing is, uh as formula then touched upon this one, whatever we are doing is everything through manual. That's just like a demo for us to understand the little bit of internals. What is happening where the connection strings are where the configurations is changing. So for that reason we are going in doing mainly technically everything will get deployed through ecd and another important point is accept the dev resource group.

C

You will not be given access the right access to any of those resource resource resource subscriptions like only the dev subscription. They might give you so that you know you can do some experiments check that and learn about this configuration only for that. You will be given the access and dev subscription from test qa stage you at and fraud. No, no one will give you an access right access. They might give you read access so that anytime, you need to check whether the configuration is correct or not.

C

You can go and do an investigation, but never ever you will get access to the the any subscription more than dev, so whatever we are doing is assuming that you are doing in dev and everything should be done in uh production manner. Okay, now that we have uh added that we did the white list and we are going into the database and, oh god, it's troubling me, like anything.

C

Today, okay, so now you see this, there are no tables, no stored procedures or nothing. So it's a pretty much a clear-cut empty database. So what we can do is pretty quickly. We have a database project here. The there are multiple ways: some people they will use a jumbo sql file and do all those changes and maintain that in the maintain that in the repository azure radio and they use this uh build pipeline and release pipeline to validate the sql scripts and then go deploy it to the to the data store.

C

So that's being uh that's being done in one of the some other people they do it and some people they do this uh sql uh project, which will be creating the backpack and deploying it so you'll see this.

C

It's a pretty uh simple thing: we have the books table and you can see that it has got few columns id and these are the few columns and we have this and there are few basic stored procedures like add a new record and get all books by get all books by get all books and get booked by id and very, very simple, uh stored project, update and delete for us uh for today, we'll be focusing only on get all books and ad book.

C

Okay, we have this one, and here is the one where I have that in my local and I'm going and deleting this one stood.

E

C

Okay or I can go ahead and that's the data store, we have this books. I want to do some kind of this thing uh droppers this one. I think I need to use the sql server. This guy doesn't give me that access okay in the meanwhile, let's proceed. So what I'll do is?

C

I will go ahead and set this as a startup project, and we have already seen that we don't have anything here in the server and and one thing as we said this as a startup project and right most click and go to this properties and from this properties uh a couple of things: the project settings, the target platform is sql, server, 2016 or 14 or 12. Whatever is appropriate if it is on the on-premise or on your local laptop.

C

You need to ensure that you, you will choose a correct version of that and then go ahead and click on start and ensure that you are having this as a startup project, and you will see that it will go ahead and compile this project and it creates the dac pack over here and it starts the deployment and woo okay. That is the second part I will come back over here. I have configured to azure sql, that's a.

C

Problem, okay, so I will go ahead and browse I'll go ahead and browse and in local I will select this sessionsdb and quickly do a connection test. The connection it is connection is succeed and then we'll go ahead and quickly do a start of this you'll see the dac pack is there and it will not take much time less than a minute. It goes and deploys that.

E

C

Creates a dac pack and also if we keep a watch on this folder, we will see a sql file getting generated. I will go ahead and open this and keep myself ready and wait for that. It ensures that the all the scripts are in a compile, compilable state and the advantages advantages. We can ensure that all the scripts are right and incrementally we can do deployment and the backpack, and this project itself will take care.

C

You can see that it just created the sql data store and it deployed all these things it has created, and I think I we don't need this one to.

C

I will go ahead and close this one, so it deployed here in this one. We can quickly come back and right most click and do a refresh and come back and check the databases. You can see that the database has been the newly created database has been there. So if you use existing database, it will go and update it accordingly, but we have selected the new one so now that we know how to do it locally.

C

So if you want to do that to sql azure all we come back over here and select microsoft, azure, sql server and then come back to the debug and click on edit and we'll go to browse and then azure and let's do a refresh. You see that sql server sql db, bookstore prod.

C

So that's the connection and I'm giving the password of this one and do a test connection. So the connection is succeeded. I just click on ok and let's start so again, it does the same uh same process.

C

We can keep a watch on this like the tables or anything getting created. So let it take a little bit of time uh for this to go ahead and do in the meanwhile. We can uh start uh discussing so some of the things. What we will learn is no problem, let it be so we'll come back and save this connection. String and I'll save this connection string here.

C

Okay, I'll save the connection string and another important point we can we can see over here is the this one: the transparent uh data, encryption encrypts, your data, backup and everything at rest. So it's sure that whatever the data which we store in sql azure, they are encrypted just that we have the credentials. We are able to see the content in clear text and another important thing is: we have the performance overview and performance recommendation also so sometimes when we write those sql stored procedures or the tables without proper indexing and other things.

C

This gives us an uh decent amount of.

E

C

That how can we improve- and this is the one feature which I really liked it- I never used it but uh reading about it uh was a pretty good thing, see see this. What it is doing is the key three important points it does is: let's assume that you, it has executed a plan, execution plan it has got and that plan is uh executing the stored pressure within two within a second less than a second.

C

Let's take it less than a second, and afterwards we created an index or it created an index or dropped an index in any combination or the content is more and now the current execution plan is taking two two and a half seconds to four seconds. So what it does is it will again it has got an intelligence to go back and forth the previous execution plan so that it gets uh faster, so it creates automatically it understands and creates the indexes and what are the indexes which are not applicable?

C

It deletes this, so these are the top three things which it does and that's that's one one important thing and another important thing. I just wanted to show you here is the one thing? Is this yeah, so this one? So another point is uh another point: is we thought about? We we told that encryption at rest. Now, what is uh what is the flexibility or a feature? We are getting it when the data is in transit, so we know that in rest, it is encrypted.

C

What is that we can do so that you know, data in transit is also secure, so we can use this feature always encrypted. So if, if you do that always encrypted data will be, I mean if you go and use the select on the table that particular column or the entire table, whatever we select as part of this one, those will be in an encrypted fashion. I mean, when you do the select query on this uh on that particular table.

C

You will not see the original content like ssn or other number, those kind of things those will be in encrypted, and even if you do a select query, it gives us an encrypted one. So we need to use a libraries to go and do that uh transformation. So I actually, I could not locate that I did one sample on this one. I could not look at but use these two features from the security perspective. So let's come back to this and all is well, so it deployed everything.

C

So, let's quickly go back to query editor, swami, anything to add from the security perspective swami on this one. We covered the v-net integration and, and we covered about the auto-tuning and we covered about the uh uh the data interest, as well as the data and transit. Anything to add from your.

E

C

Yeah yeah, that's that's pretty much. We we covered everything. uh So uh the first piece is down. uh I think we have uh uh had this one. So let's go and talk about this uh web api and I'm closing this and anyways. We are going to delete this database after this demo. I am letting my secret password know.

C

Okay, so that's my.

A

A

How we are securing these current sales and everything exactly.

F

C

A

C

Right, so what we uh thought uh to do is we thought of to give a short on dot, net 5 and dot net 6, and we didn't notice any uh considerable difference, but uh the same piece of code. We were able to use it, and here I just wanted to highlight one piece on the reusability, so we created set of dlls uh libraries, which is this data store, bll cash, dal, core and sql, like it's, a mix of n layer approach. Come the domain driven descent, so the.

E

C

Namings was the mix of both the things and just wanted to highlight one small piece over here.

C

uh If you see that swami correct me, if I'm uh going off the track so see that we have the set of uh front ends, blazer was a manga reactor view js or any number of things are here in the game and the data store. We have it and this middle tier and there's a people are using it interchangeably. Someone will call this as a back end. Someone is calling is apa layer or middle tier whatever it is, so I'm calling it as a middle tier if it is wrong.

C

Apologies for that and uh inside inside this tier we are creating api, so we'll come back to grpc uh in future, but as of today, today's demo is with the api, and that goes and talks to the business logic layer and you have a data, access, layer and core. So within the tier we have. Logically, we have different grouped the code to different layers, okay, so coming back to the code, so we have all this one. So why do we need to move one of the advantages?

C

You see that if I we have created this uh in two different frameworks, one is dot net five and dot net six. So what does this api project contains? Nothing. It just contains the startup cs and one controller, and what is this dotnet six one contains is one startup and one controller. So where does the code lives? The code lives in this this libraries and tomorrow. If someone comes and says that hey um I'm going to write a grpc, one of the customer is asking for an integration, and what does he does?

C

He does not do anything. You don't need to reinvent the bill. All you need to go and look into these contracts, create the create the protofile and write a service and reuse the same components. Nothing will change, we'll show that benefit in future episode, but for today you see that benefit all the like. All the code has been pushed to the libraries. The benefit is, we are not repeating ourselves anywhere.

C

Only thing is we are looking at uh hooking it up in the um in this respective projects. So, coming back to this startup, so in this startup we have white opened up pretty wide to the entire world, any origin, any other, any matter any method plus please please do not do this. This is only for the demo and because we will be adding multiple things coming from a different locally. We need to test and a lot of places just wanted to keep the changes minimum.

C

We just opened it for the while, but please do not ever do that, and another thing is how we are hooking up. This uh jwd token. Well, we'll come back and look into look into that a little later, and these are some of the dependencies we have application dependencies and another important point is: we have added the use authentication in this okay, now, coming back to controller, we have added the authorized, so only the authorized person will be able to look into this and it has got pretty simple uh methods. It is.

C

It has got a post to add a book and uh get to retrieve all the books and uh get by id it. We will get this and uh we will come back and we'll start uh executing this step by step and we'll deploy. Also pretty quick uh see that I'll I'll see that we are, we are doing it uh step by step. Let me go to this app settings here and yeah. So app settings- and here is the the connection string here- is the connect system. I just wanted to highlight on the cache side pattern.

C

So that is the reason I'm stopping in here. So right.

E

C

Oh okay, so this is the radius cache on my local. um So you do this uh keys. There is nothing stored in this. uh Nothing is stored in this this one. So let me go ahead and add this.

C

Okay and that and coming back to.

E

E

D

One minute.

C

That's an api. Let me go.

C

To the kudu console.

C

Technically, I don't need this kudo console.

C

I can do to go to the kudu console. This is an existing one, I'm showing, but we are going to spin up our own in a moment.

C

So we can come to this kudu console anyone who is pretty good in console. We can use this and I can go ahead and check this. You can see all those connection strings and everything over here or I can go to the console, and here we will come to know that how our application has been deployed. It will be on the site and dub dub route.

C

Okay right now, oh there's a broad subscription. I'm sorry.

C

We don't see anything over there because I went into the plot subscription.

C

Let me go either way you can go to the prod subscription, I mean go to the kudu console or the app files just give it a moment, and the difference between the fqdn is that sem will be added here and you can come in here. You can see the things whatever as a configuration which gets added over here as swami was right uh pointing out about this one, even though we test it and we deploy whatever someone has access to this resource or a resource group through our back or whatever the mechanism.

C

Even they go into this particular particular.

E

C

They will not be able to get any credentials because we are not at all storing any credentials and that we will store only in our local as as long as we want to test something we'll use that one. Okay and another key thing is we come to configuration, so this guy has got access and is trying to find out some pieces of information, so he came to this file. He didn't find anything. So the next thing he knows is okay. I need to go check here. Oh it's here.

C

Let's see what it is, so you click on that, and here also there is uh nothing in here all it is pointing to a key vault, okay or zero authority, so who in this world will be given access to keyword very, very few people? Only the devops lead our architects.

C

Only those kind of people they'll be given access to this one, and also that too, based on the policies I mean, if you see that this application api application and that to what permissions we have given, we have given only this get and list technically it should be get. We don't need need to give this list, but I have bumped it up, but I I have a lot of lot of permissions on that one.

C

So if you come to this secrets here, um if you come in technically you will not be able to come to this level because they will never ever give you. uh They will never ever give you this permission for uh permission for this for sure. Okay, let me go ahead and copy this bookstore.

C

I'm keeping that that's the audience, I'm keeping it okay, so I just wanted to show that piece about the cache aside pattern, uh we'll do very quick things and I have I'll keep one debugger over here. So another important point is throughout our project. If you see that lot of things are getting dependency injection, so in this books controller it has got dependency on the book's bail. So when you go to this books, bll, maybe the naming convention I messed up, but it's a mix of the domain driven design and the layered approach.

C

I just used both approach. I am, I know, I'm wrong, but that's how I did it so here we have the books repository and has a dependency on the cache repository and I logger. So these all things are coming in here, so someone can ask for me. You have done and I'm seeing that the application dependencies has been hooked up over here. But where does this logger come from? Who is giving this?

C

So that's what we'll see that in a moment and then we will place the debugger here in the get all just wanted to show you one small piece. So let's stop over here. We will take real three minutes time to go over this and then, let's.

C

C

Okay, so this is the one which is uh coming in at this point of time and swami. I might go on mute because sometimes my laptop, when I'm going on this quick watch, it mutes my mic. So please give me a shout so when you come and see that before even hitting the first line, there are 79 uh dependencies and you can get into this result view to see that what all those default dependencies I'm calling it out of the box dependencies or default dependencies.

C

So you can, if you look into that, I logger is one of that I logger is one of the dependencies. So now we understand, wherever you are using this high logger, it is coming as part of the default dependencies or dependencies coming out of the box and and other dependencies which we are adding just to save the time. I'm just hitting this one and let me do an f10 and at this point of time, if I go and uh do a quick watch again, I see it was 79.

C

Now it is 245 dependencies and we can expand this uh result view to quickly go over that all the dependencies which has been added so yeah. You can see that, apart from all the dependencies which was added our uh application dependencies, you see this ibook repository, I cache and whatever we added of we added from the application dependency side, you see that the idata store connection strings and reduce everything whatever we added it. Everything will be here, so this is like a container, so that means it holds the entire dependencies.

C

It's like you have deposited money in a bank account, take the credit card, go to an atm and swipe it. So if you don't load your dependency inside the dependency injection container and anywhere, you ask it like in atm, it says the decline transaction, your the application will blow in front of your eyes.

C

Okay, so let's go and quickly do this uh one end point: uh let me go and grab this guy use the postman I'll come in here and drop this slash api, api, slash books and we'll do this because we are not sending the token it will definitely blow.

C

I will go ahead and it does add.

C

Authorization and have this now, I'm sending the request, so we have sent the authorization token. You see that first it comes and hits the key store, um but it is there's nothing in here, so it goes and finds out. There's nothing in here the string is empty. I mean there might be better ways to do it. This is how I implemented so then we are going to the sequel. So only when the data doesn't exist in the redis cache, then I'm going to the sql bringing the contents.

C

Now we have the books now I am storing that inside the reduce cache. So I do continue so this point of time, if I come and I uh see the keys, there are uh all books.

C

Now so it is it's there, so it is my duty if someone adds a new record. If someone adds a new record, it is my duty to go ahead and delete it. So if not people are using the scale data, they are not seeing the latest data. Let's do that real quick put it in a json. I will mithun.

C

Nair is my best friend so 30, and I'm deleting that so here what will happen is now we have to come and see this when someone is going and adding a record, so he adds the record and then we will see that we will go ahead and remove this from here.

C

Okay, so we'll go and grab this path come back over here, put it over here and headers. We need an authorization.

C

Let me grab that token from there I'll put it over here and we have the body and let me send this record and it comes and it it has added a new record and it at this point of time, you're, seeing that the radius cache.

C

Redis cache has got this entry. Let's do this one, and now you see that that entry is removed, because now the single source of truth is sql server, because there were only 10 records in the cache and we added in 11th record in the data store sql. Now there is a difference. The single source of truth is sql, so we need to invalidate the cache. So we did the invalidation of the cache and we did a continue and the next cycle.

C

Again it goes on okay, so I just I think, swami did we covered the cache's head pattern in a decent extent, or did we miss anything anything to add from your site.

A

I think you're pretty good uh swami yeah.

C

Yeah, so the situation is people call swami me also, and you also that's uh when we're.

E

Calling swami, we don't know.

C

Whether we are calling I'm calling you or I'm calling myself, that's one confusion, okay, so uh pretty much. We covered a lot of things over here, so we have the connection strings and all those things. So let me go and close all the tabs just to be because I thought it will be easy to showcase it here on my local instead of on the server okay I'll leave this over here.

C

Let me come back to home and go to the broad resource group and inside the broad resource group. We have the key vault okay. So let's start having those key vaults and here is the secrets, so access policy- if you see that only I have the access to read, even if swaminathan logs in he'll not be able to technically not be able to do that view it because uh I have not given access to anyone.

C

Okay. So let's come back to this one and I will go to just to save some time or avoid the clashes, whether I'm able to giving this I'm giving the same name or what?

C

For that reason, I'm coming back to here and I want to use the same same name so sql server connection string, I'm just creating the create a manual and create manual, and where is the connection string.

C

So the connection string- this is the one windows.

C

Yeah this one broad, okay, I'm coming and I'm see that it has been masked right now it is masked so I can go to create it creates it, and even if someone wants to come and use this one here, I need to speak about one important point. So this is just a test.

C

I'm adding some content, and I just created it so there are one important uh setting which you can select is purge.

C

Protection, earth protection- I'm very bad in documentation, so I rely on uh rely on the companies to use the content. So if you, what is the difference between soft delete and purge protection? So soft delete? Is this one soft delete is enabled by uh by default? Okay. So what happen is if you, the soft delete, is enabled. So if someone is coming in and their resource, they are deleting this resource.

C

So what happen is that will be put into uh to be deleted condition and it will be there for some time so maybe 7 days or 90 days, whatever is applicable, based on that it will be there on that on that bench for us to go ahead and recover that okay, so, let's see in practical action, uh I have this test, I'm going into this test, and I'm deleting that okay, so that's uh it is deleting and it didn't delete. We have to wait for some time in the meanwhile, let's discuss the siri.

C

So what happen is once you go to this manage deleted secrets? You will see that all the items which has been deleted over here so now you have two options. You can recover that or you can purge it so before you do that, whether you'll be able to purge it. You have to come and see that here the purge option that privilege is not enabled. I mean I didn't, give that permission even to myself, because the prop the problem is it's a double-edged sword.

C

That means I have. I have created this and I have deleted it and I will be able to go and look into this manage deleted items, and I can I can select that and I can purge it. That means it goes out of this place also, so no one will be able to imagine a scenario for some reason. I want to trouble swaminathan and he has given me an access to his resource group.

C

I can pretty much go and delete that and come here and purge it and uh he has got a vacation on friday and his boss will call and say that swami this is not working. Can you take off next week and work over the weekend to fix it? So to avoid those kind of scenarios, I mean wantedly or unwantedly. For any reason, if you someone deleted it, they should not be able to purchase. There are multiple levels of security.

C

First security we have already seen that will not be able to give this purge to anyone, but even in in worst case scenario that privilege is there with me. I will be able to go ahead and purchase, so microsoft has covered that kind of scenarios also. They said the purge protection.

C

That means, even if someone deletes that, even if someone deletes a key and if we enable the purge protection. So what happen is when you enable the purge protection, they will ensure that seven to nine seven to ninety days, the configurable, if you say that whenever by default soft delete, is enabled. That means you delete any of this keyword item. It stays on the bench for 7 to 90 days.

C

Whatever is configured and within this 90 days no one will have permission to go ahead and purchase, even though they have the access policy, even though they have the access policy. If the purge protection is enabled, they will not be able to delete it. Okay, that's one thing I just wanted to cover: that's a pretty important security feature we have to enable it might come with little bit of cost, but I think that should be fine um yeah. We are good now. So let me come back.

C

Anyways, I'm going to delete this also from the r0. It's all for the demo. It will be there for only next four hours or two hours.

C

Okay, so let's go ahead and do a deployment. We have already seen that we already seen that. Let me come back over here and go to the prod.

C

And go to this api.

C

Go to this api and I just want to show the harder way so I apologize if I am troubling you people, but I want to. We can see this app service editor also text editor, but I want to uh show the hard way to do it: how to do that, so that's kudu console. We have seen that when we come to this, you can see.

C

What yeah you can see this site will get deployed here, there's nothing in here. So this is the hosting start. You see this is the page which gets deployed if someone goes to this end point. If someone goes to that end point the whatever the hosting is hosting html. You see here hosting start html, that's the html it gets displayed. I mean we have used the shared app service plan, so it will be little low on the resources. Apologies for that slowness.

C

And uh you can see this environment is very important thing where you can uh it's like a keys of the kingdom. Anyone who has got access to this one. They will be able to see everything over here. What configuration is there so be little attentive to this one? So, okay, that's just an fai. So this is the file.

C

Let's go and deploy.

C

You do publish, and again we just want to call it out that whatever we are doing is just to know the internals are a little bit more knowledge. Definitely the devops team will not allow you to deploy directly from your uh visual studio, visual studio code or as your cli route or the mechanism for sure.

C

So let me go and deploy this. I don't need to do that. I can do the publish here and.

D

Hold on, I did the mistake.

A

You didn't change this group or something.

D

Yeah yeah yeah yeah, oh yeah. I did a mistake. I.

A

Didn't realize that no no no race, so, yes, I think we are close to an r in our session. So I just wanted us to. You know: take a bit of a pass.

D

A

Answer a couple of questions, and then we could. We could continue forward.

D

A

Yeah, you can continue deploying it.

D

A

Just click on this reply button and then maybe yeah.

D

We can, we can take those.

A

A

I also have some work for the audience now to make it a little more interactive, so I'll just put.

E

Out once you are.

A

C

E

E

E

A

C

Okay, that one and other things we will set it there itself. Okay, let me go ahead and click it, because this is the one which everyone wants to avoid. I might do some mistake over here and site will not work, and everyone will come on to the bridge to hunt this down. It's an unnecessary uh pain and waste of efforts, so cicd will save us from this. Okay, I'll stop sharing, and while that is uh okay, we can keep it.

A

Sharing I'll come.

C

Yeah, you can continue sharing I'll bring my.

A

Screen to the stream, okay,.

D

A

So I would be posting a link in the youtube chat window, so I would request all the audience to just go ahead and give you a comment. Whatever will be good bad, actually, don't worry about it. We just wanted to seek your feedback over the course of the session as well, and also to make you do some typing right so that you don't really feel the monotonous sitting or for the last couple of hours or so so. Okay,.

C

Swami I'm sending it some urls also.

D

C

The soft delete and purge protection is very important.

A

huh Sorry I had a momentary glitch, um so this is a request to the audience. So I've picked a link like a mentee.com link, I'm also showing up in the notes here it's available in the youtube chat window. Can you please go there and then key in the code? Whatever is shown in the in the stream now 9073817 and then you can just give your feedback there and then it just comes up here. Whoever wants to take a break for five minutes or so please, if you do, that, whoever is there in the stream can.

A

If you can go ahead- and you know key in your comments that will be good, just go to.

E

The site mentee.com.

A

And then just use the code.

A

Eight one: this is that's one of the ways to keep yourself more engaged because in a virtual environment, it's pretty difficult to understand the audience uh responses, and things like that, so yeah we just wanted to all right live demo is good. Thank you good to know that people are following us.

A

Okay, as as that comes up, maybe I'll bring a couple of questions which one then we could try answering them. Okay, yeah sure sure. First one is around your backpack thing, uh which is you know which visual studio extension is used to run the sql script. I think you show it directly from the vs and deploy it into the you're right. Do you want to show that explain which visual extension are you using? Is it out of the box thing or or why.

C

Do you think it uh uh okay, one second I'll show you that uh am I sharing my screen.

A

Yes, I think you are sitting back this.

C

Yeah just bring bring my screen uh yeah, you are there, so this is the uh sql proj uh visual studio 2019.. So the first thing is this: one I'll give this url. Also, so you guys can see this one, so sql server data tools. This is where you need to use it.

C

So once you use that one, so you will have the project template so where you can go and right most click and add new project and inside this template you will have that sql project and if it doesn't, we can search it I'll show you how to do that. So you can see that sql server database project this one.

A

Okay, okay, excellent: I hope that answers.

C

So I have one backup plan also if anyone asks the toughest question I'll jump out of the call and say that I lost the internet connection, I have a backup plan. I.

A

Hope you didn't have to do that. Okay, so.

E

There is another.

A

Question which I'll just bring it up now? Is there any better approach to update or invalidate the cache from sql? Like events based approach? Will.

B

A

What we typically do is the the application or the api which is going to write the data into the into the db could always update the cache or refresh the cache. Also that's one way other way is: if you kind of know, what is the the data expiry or sort? You could also set an expiry time to it and it gets automatically expired, and then you could do the rights later to it. These are couple of options which are available and yeah, which way you could add.

E

A

Hey you're on mute, we're not able to hear you.

C

Oh I'm sorry, I was talking on mute okay, so this is my favorite topic. So the thing is that uh what the first consideration is, what kind of content you want to bring it on to the reduced cash? So the first one is, uh in this case books how many times once a books get added how many times that books get updated the same book very rarely throughout the lifetime it we might say that the book is active or inactive, but that that too, in a lifetime.

C

Someone published a book this year, so for next ten years or five years or minimum two years, it will not change until unless that goes, the books goes out of this one. So those kind of things and like if, if professors are students in a college, very rarely they'll, their data gets updated either they're changing the address or professor got a raise in his paycheck, or he has got a promotion very rarely so choosing the data.

C

What you want to store is the first important consideration before before going ahead and using the radius cache, and the second important point is how uh second second important point is how much data you want to bring in, because this is the uh in memory. It stores the content in in memory. So you have, the pricing tier will go into the pricing tier of that reduce cash. I have already placed it yeah you. This is the one which I'm flashing it here so in this pricing tier.

C

If you see that the basic you don't have the availability, nothing is there. We are using this one. So if you use this, uh if you use the standard one, so they are giving the 250 mb, plus 53 gb and they have a replication. The failover is there.

C

So if it is a premium, so they are giving 6 gb uh to 1 120 gb and you have the failure, failover and some of the other features, and what is pretty important is we are getting the size whatever is the size and as in when you grow as and when your demands are more, what you can do is you can start spin up a cluster to bump up this right now you have only 53 cash size, whatever is the cash size at 53, and this one is 120 gb and what you can do is you can spin up a cluster and add more of this ready cash so that you will be able to access it?

C

The third important point is how much time this content should be there on the radius cache. So that's a pretty important thing. So, as swaminathan already pointed out, one of the thing is: if you want to automatically remove from the cache you can do that uh tag it the timeline time to leave, and another thing is, uh I personally don't prefer the time to live, because it is our responsibility um if I say that use this set of books for 10 minutes.

C

So are you sure that there is no other book is getting added to the database within 10 minutes? I'm not sure at least so. In that case, what is happening is the difference. What is there in the redis cache and what is there in the sql server doesn't match, so whoever is using the cache, because, according to our code, it goes and first hits if it is a hit, it gives this content from the cache.

C

So there is, there is a difference, so single source of truth has got 11 records and the cash is giving you 10 records. So.

A

I mean, isn't it.

C

It's just to add sorry.

A

D

It also depends.

A

On the use case right, you can't always use detail, or you can't always use even based. Also, it has to be a mix of mix up of things right. You need to choose what is the data set and how frequently it changes and then how long it needs to be lived, etc. Right, if, let's say, if you're caching, a data which is not your own transaction ladder, but from an external system that could maybe live for longer right if it's not going to be changed very often right.

A

So caching, strategy by itself is a bigger topic all right, so we need to choose. What is the data set? How big is the data set and how long the data should should be there and how? How frequently it should be refreshed but to factor in all of these things and then come up with the caching and the invalidation of cache, etc. Yeah.

C

One final point: the one final point for uh prasad is: you can very well do with the event mechanism also, and that depends upon the moment uh here. What is happening is the web. Api itself is taking a decision whether to remove the content from the cache or not. Imagine a serial scenario like you placed an order and they say that we received your order and right now they are not confirming that the order has ordered they have taken it or not. They will put it in a queue.

C

There might be some worker processing which takes this message from the queue and adds it into the database after verifying your history, whether you are the regular pa, uh you are paying it regularly or there are a lot of check bounces from you or different scenarios, then either they will add the add this record into the table table so that you they will ship the book or they will reject your order. So in that case, what happen is the moment?

C

They add it, so they will also need to add drop a message in some other which a process exclusively works only for this, but only for this removing the content and refreshing with the latest content that will pick up the message from the queue and gets all the latest records and drops it. So that's a little bit of more work. So me am I uh on the right track on this one like order.

A

Fulfillment kind.

C

A

Yeah yeah, I think yeah that's again on an even driven architecture, kind of right so that that basically.

D

E

A

E

A

Handle the entire micro services and even driven fashion right, okay, getting back to the topic. I know there are a lot of interesting questions coming up, so maybe we'll take next 15 to 20. Minutes to you know close the demo part and then we'll have a q one day, starting from 12, five or so. Okay, maybe we'll have we'll take more questions by then yeah. Oh.

D

A

D

A

I think we are getting good responses, so I think the content and the demo is going good, so you could continue.

C

I would like to get any negative. I would like to hear any.

A

Negative response or a response which is lower, I won't say, negative, some constructive feedback for us right, which is like which I completely agree. That's one of the reasons why you try to pull this in now, but yeah with the virtual sessions. There are little other technical challenges too, but let's see how we can make it more interactive.

C

Do we have any constructive.

A

Here it is, it needs to be more interactive, is one of the things which I could see: okay, otherwise it's an one-way one-way, dumb from the speakers right, not able to hear from the audience yeah. Okay, we'll try to address them differently, yeah sure sure. So let.

C

Me I'll bring.

A

Back your screen and then maybe I could continue with that. So.

C

How much time do I have to spin up the entire uh app? I will cut down on this piece by piece and I'll go ahead and set it up. Everything.

A

Yeah, maybe another 20 minutes, you can take it 11 45 now, maybe we'll go till 12 10 and then we will we'll take more questions and then we'll answer them make.

C

Sense make sense, so let me quickly go ahead and deploy the web. Also, then I'll set it up the key vault and everything and we'll see one round of uh yeah that that makes more sense and let's.

E

C

Yeah yeah, so we can talk while I'm doing this so and if you have any one or two queries which we can discuss while I'm deploying we can do that. One also.

C

And this app service itself. I have piled up lot of things to explain, but I don't know whether I'll have a time or not, but I'll see what best can be done.

A

Yeah yeah right so yeah since you're doing on that, I just wanted to pull this up right around the azure keyword services. Right I mean it's not uh definitely it comes with the cost, but it's not a huge cost. It's very, very minimal. That's one thing and also well. The key walls would have tightened iterations. I think amish also pointed out in the comments right akb, which is azure keyword, has great integrations with all other azure services. The tab, service or kubernetes service functions are variable right.

A

It's it's one, common stop shop for you to store all your secrets, keys, certificates, etc. Right, if you want to move to another cloud, for example, all you would need to do is just maybe take these secrets and then put put it into the cloud native cloud. Specific secret store there right in aws, so other option is the photo you are using kubernetes. Yes, like you mentioned, like I'm just mentioned, you can use either k8 secrets or, if you want to keep it completely cloud diagnostic, you could also go with the third party.

A

uh Walls like hashicorp walls or different other walls are available, which would become like an external to the cloud, but the thing is that tighter integration would not be there. What I mean by tighter integration is now when he shows the demo right. He would deploy the dotnet app into the azure app service and from app service. You can directly connect to the keyword and then you can pull all the secrets during runtime or during build time, and those kind of tight integrations would would not be possible if you are using external external walls.

A

That's the only thing, but otherwise it's not a hard and fast rule to go with keyword. But if you do it, you will get other benefits. Also.

A

Okay, yeah, please go ahead. I.

C

Just want to highlight I just I just want to highlight one piece over here is: ah I need to look into the keyboard, I'm I'm in the wrong place. Apologize for.

E

C

I just want to bring one important point here, this one, so this security comes with uh backed up with uh hsm hardware security model. uh That means I I worked on. I worked on on etm applications. So what happen is the hsm security? Is that it generates this keys or 32.

C

I'm talking I'm telling this story about 2010., so the security, whatever it generates out of this hsm. That will be in two different pieces. So I know the first 16 digits of the key and swaminathan knows the second 16 digits of the key, so he comes and he keys in 16 digits and I come and I uh key my second 16 digits, so that's secured. So what is uh hsm managed? You can have uh use this one. This is a pretty much tightly secured.

C

The hardware security model which generates the keys and a lot of things are there. I mean it's very tough uh to crack this one, because until unless both the people are using that you can't crack it, because only you will have only the partial key and the second thing is, you can have your own, uh you don't want to use this. Hsm security backed up and you want to bring in your own keys. You can do that one also.

C

This is also highly secured that one point I just wanted to bring in while it is getting deployed. I just want to speak on this app service. Also. This is pretty uh pretty good thing. So, while you take this, what is available so if you use free, you will get, you can deploy, 10 websites 1gb and it's a shared compute, so you will have 60 minutes of dedicated time and the shade you have this thing.

C

The only thing is, you can support the custom domain and here in basic you can scale, but that scaling is manual and when it comes to standard you have space and unlimited apps and uh you can do uh scale up to ten. Here starts the life of auto scaling.

C

So if for for some reason and- and it goes on like this- and something known as isolated, that means you are not using the uh here. It is a shared compute and here is a dedicated compute. This is an isolated. I mean the dedicated compute is other people also can be there within the same infrastructure, but here it is only you, no other person will be sharing it.

C

Okay and another important point I just wanted to use is this authentication mechanism. I think I will not have some time. I just picked it up, but I uh will go into this uh settings section so in this settings section there are a couple of important points. I want to highlight that the key thing is, you can do a bulk edit or- and this is the key thing so in our configuration in our local configuration. If you come and see how do we use the connection connection strings or whatever the settings?

C

I said that odd zero authority, r0 colon authority, r0 colon client, id r0, colon response type- so that's how we go ahead and read it from within the from within our code. This is how we read it. Web api is colon book and uh and rest of the things also, you can see this baby colon books, so that will not work when it gets deployed to azure. So what happen is this colon has to be replaced with double underscore underscore underscore?

C

Please ensure that you're doing that, and you can do it uh bulk edit, we'll see that underscore underscore in a moment, and another important point which I want to highlight is this will come into the picture when you're using the connection strings from this block. I would quickly switch over here.

C

To this guy I mean technically, I just want to give that right. Now, I'm not using that way. Maybe in the next month's demo, when we introduce azure function, we'll definitely I'll definitely show that the this we will come back to that one. So if you have the connection string, so the block itself will be connection strings, colon, something default connection string or something. So if I come back over.

C

Here yeah here you can see this so the connection, the connection strings, colon connection string because it's a child object in that json. So what happen is uh you should be very careful when you are using in azure functions when you create the uh when you create this uh sql, when you create this uh connection string for azure, so it will prefix this with this uh sql azure con str underscore, and if you try to access like access the way which we do it locally connection strings colon, then it will blow.

C

So you need to ensure that your the way you are accessing will also change. I have used in the core, I just created a so if it is like a connection, strings, sql connection string, this will work here in app service. When you go to that azure functions, it should be the. It should be this model, so sql azure con str underscore your sql connection string. If not, it will be gone. You'll spend hours together. I spent myself spend four hours to crack that it was my mistake.

C

uh So that's one point and the second important point to highlight over here. There are a lot of things to discuss, but I'm just highlighting the one which troubled me, so I think that's it from there and I piled up few important points to discuss. Yeah scaling is very, very important. So how do you? How do you scale?

C

Okay, I mean scale up- is adding more resources and scale out is adding more instances? Okay, so here we need to come and verify that what uh app service plan supports. What facilities features include like you say that I want a pizza with pineapple topping you say: that's 20, rupees more with another topping you'll say this are 40 minutes, 40, rupees more! It's like the same toppings.

C

Okay. That shows that I'm feeling angry I'm talking about this one. Okay, so uh the scaling is very, very important. um The configuration we have gone through the configuration yes. So this is the important point which I want to highlight here. So what we can do is we can scale if it is the basic one you can do a manual scale and if it is standard and double you can do a autoscale so in auto scale.

C

Also, there are a couple of things which we can uh do so for the first thing is where, where you, you have scale based on the schedule, so what I'm telling is um next month, like uh whatever the festival, whichever is coming so on those first 12 days.

C

We want to bump it up to three instances or two instances, so we can say that uh our on weekends reduce to one sentence, uh one instances, and you can say that okay, minimum maximum and the default so based on this scaling and scale out and scale on schedule or scale differently, on specific dates, like whatever the festival like black friday or whatever, not coming in number.

C

So sales will be huge, so we can increase bump it up to more number of things and- uh and another important point is we have to auto scale based on the metrics.

C

So, whenever the auto scale, based on the metrics, whenever the cpu is more than 70 for past 15 minutes or 10 minutes, bump up a new instance and while bumping up the new instance also what we can, what we have to definitely do is we need to ensure that the load is because, while the new instance is uh coming the current instance or the current instance, whatever is whatever is accepting the request, it should not uh blow because if it is more number of users are adding and it goes to 75 18 90 and at some time it might say that it's not able to process or it's taking longer time or it might say, 503, that's one consideration and another important thing.

C

What we have to do is use the slots staging slots and ensure that your swapping based on the staging slots- I just want to you- show that one important point.

E

C

I'll close this one so I'll get some extra space.

C

As azure okay, so this is very very important. uh While we are switching between slots, so you can see this. We can add a slot and test it like you have staging and fraud and we will switch while switching also, what we have to ensure is. We are doing a balance of this uh traffic percentage. So how do you want to handle this? Do you want to send 100 of the traffic to the new and uh nothing to the old slot, or how do you want to 50 50 60 40?

C

Whatever is the combination with real time scenarios, because the new slot will definitely they uh they see that change and another important point is: please go through this documentation, so we need to have that warm-up and application initialization. So it will ensure that uh until unless the application installation is done in the new slot and the warm-up, that means it is able to hit that end point and it is able to get the response.

C

This is a pretty important point which you have to consider before you do the auto swap- and here you are saying that until unless I get a response code of 200, don't send the request to that one okay, so these are pretty pretty important things which you have to consider.

C

I think this is done, so I will quickly go and set up our environment, real, quick, I'll, close everything there might be some glitches, but let's hit the road first. This is keyword and I'll come back to this one. This is our prod, reduce connection string.

C

I'm coming to this one.

C

C

I'll go back to the prod instance, get the radius and there are again some consideration in the redis cache. Also um redis cache is in memory. So if some piece of uh some piece of information which is which you are storing in red, is that is in memory, someone can get that information so.

C

Security, so you need to bump up your radius cache to premium, so that will be backed up with the content, storing in the storage and where you will get the benefit of benefit of data in rest, so somewhere it will be here in this. uh You need to go and check the uh security practices over here uh you, you can have this uh network house isolation. This is one thing and another. Another important thing is the storage address. Rest, that's very important. Let's see that encryptions are in encryption yeah. So this is.

C

This is one of the important things, so you need to ensure that how much it is there and what uh level of plan you need to take so that you know you will have this. This is the one which is there, but the content which is stored on the memory. If someone's uh get into that it might be dangerous. So we have to use that security practice. While I was creating it, I got the.

C

I was able to recall that okay, so always use the that's my personal choice. I might be wrong, but I always use the uh secondary keys only. I never use the first primary case for some reason I don't know I always use the secondary one.

A

Okay, there's just a personal preference, there's nothing.

C

A

C

Because I was long ago, maybe 2013 or somewhere, I was reading a book and they told that it will be easy to refresh the keys if you use the secondary first go back and refresh the primary and start using primary. For some reason it went into my nerves. I don't know why.

D

A disconnection string.

C

It's a lot of things. I do it's purely my personal choice.

A

It's not the best practice and quick time check. I think you may need to fast stack a little bit.

C

Yeah I'll do I'll do that real, quick.

A

I think we are ready to showcase the uh blazer wisem and the auto part, maybe yeah.

C

A

Through that, with the existing app, that would be good.

C

Okay, then we want to make this as a logical closure and go through that piece.

A

Yeah, we could do that too.

C

Okay, so uh just uh you want to take a audience puller, let's go ahead, and do that. That's also fine with me.

A

I'm just looking at the time, so we have 10 minutes if you're able to run through it in 10 minutes we'll go.

C

No 10 minutes will be less yeah unrealistic. Like I'm preparing for failure. I don't want to do that. Okay, I can explain that there itself. uh uh That's that's good thing, let's go ahead and do it uh so folks, let's we are falling back on the demo which we set it up. uh That way. You know we'll uh understand this one and the.

A

Only couple of things I'm sorry.

C

Before uh yeah yeah.

E

One thing right: so the other blazer.

A

Is also and dotnet app, so that's also different into the app service right. So how which was showed now about deploying the web api into the app azure app service, it's the same, exactly the same pattern. What would be used for that also, but there's no difference. That's one of the uh one of the best things about using dot net right. You can. You can use the same language in the same deployments at the same development and everything, and then you can. You can do front end and the back end also right.

A

So what we would do now is. uh We will just go ahead and show the application and then show how the authorization everything is work. And then maybe you could just touch upon the uh zero integration here. How we have done yeah.

C

Right right, I just want to highlight one uh piece of information here. That's my favorite tool, kudu.

E

C

I can't resist myself going to this kudo console.

C

Okay, this is a very good place for us to find. Sometimes what happen is uh we have a scenario where the application is failing and you are getting. You are having no clue why it is failing seriously, no clue why it is failing. So you have to come to this kudu console get into that log files and there'll be different errors like a detailed errors or a lot of other logs or application inserts or application logs. So I really fall back on this guy detailers I'll show you why and also this is iis logs.

C

You can see that uh this one is logs and uh this is uh http, and this is uh application logs you're having the diagnostics and all those things. So, coming back to this detail, detailed error- I want to highlight one piece: is this one? I will want to not this one.

C

Whatever is our uh event lock yeah, this is the one I was searching for. Whatever is our windows event lock that will be here so anytime. You landed in a critical situation where you you're not able to crack it. So please fall back on this guy. This guy's gives you very, very detailed information, so whatever you see in your uh in your event, log- that's what is this so see this? This one keyboard not supported, so we will not have this clues so why this unhandled exception is being thrown or uh whatever it is.

C

So, please fall back on this one and okay, and also you can see the process explorer on here, and this logs is very, very important point uh and you see this process explorer the site extensions.

C

So that's two that shows that what commands are being run- and there are a couple of tools which you can uh use and uh another important point I want to highlight here is: I mean- which technically do with this app service editor. Also that's more. If you're a command line guy, then you can come in here and this is under wrote.

C

You will have the website, and here you can see that the times and your entire things are here and if you want to go and check what is for some reason, connection string is not working or whatever the reason. So we can come to that app settings. I can do.

D

C

Into the cat of this one- or I can select this and uh click on this, I will know what information is there. This is the important point I just wanted to show that and from the auth zero perspective. I really like this r0, that's a fantastic application, they're giving it for free and we can use it for any number of days. So what we did is we have created a application here. We called it as a this one, bookstore blazer vasm and inside that bookstore blazer version. We have this bookstore blazer version. We have this.

C

We have this here, so allowed callbacks your callback url and allowed log out, and they did a very fantastic job of creating the documentation. So we don't need to go speak to anyone ask. Did you configure how to configure? No, nothing, nothing is needed. uh I don't need it I'll. Show you a couple of uh urls, it's pretty pretty extensive, so you can come in here and how to integrate this or how to secure your blazer basm or how to secure your web api. They have given step by step instruction.

C

Okay, so you can see this.

C

That's that is, for the dot net file. Okay, the sdk is also latest one. So what we have to do is uh technically we need to come and create. This is the one which is showing that how they are accessing without authentication mechanism and and then what we need to do is we need to go to that uh zero dashboard and create an application and select the single page, spa, okay and then we need to add the uh we need to settings. We need to add this allowed callback urls and allowed log out urls.

C

So technically, that's what we need to do here so here in this callback urls and we need to add the logout urls and because there are multiple things you need to set the return to url. If you don't what happen is it will return to the first end point. So whatever is the first endpoint? Is this one? It will always go. Go back to this one, even if I am accessing from my local. It always go. Go back to this one, so you need to ensure that you are hooking up this return tool.

C

That's the first piece of information, so adding the support here in this. What we have to do is we can directly go on our live site uh to see those settings.

C

The settings I will go to yeah, I actually, I did one short circuit technically, I should have uh technically. I should have uh used a api method to bring in this configuration, because this is not safe, safe. It is, it is going to spit out all these credentials openly, so it is dangerous. I am it was my mistake. I didn't uh do a call to key vault and bring in those credentials through api and use it.

C

So I did a short of that. I want to show that so that's the first piece of information. We need to have the authority and client id, so this is the other way. Instead of using the kudu console, you can use this one. Also, the prod json is one so your authority and client id. So these two pieces. You need to have this one in the blazer app this one. Okay and then we need to add this package, and then you have to add this authentication mechanism.

C

There was one small catch which I would like to show.

C

I mean that was also given in a different place, but I had to look out for that. So in this program.cs we will add the authentication this piece, and the second important piece is where you are calling the web api. We need to add this piece of information so that you know it will pass the access token. It doesn't use the identity token. So when I was doing that mistake, I came to know that I should not use the id token when I'm doing a handshake with the api we'll see that in in live.

C

So that's the one thing which you need to do it: okay and then the next piece. What we, what we have to do is uh coming back to this shared. We have this access control. uh We need to create this component. Everything is has been documented. Nothing is missing, so everything is documented, makito monkey. If you copy, follow in the instruction in this r0 documentation, you will be able to set it up. Everything within one or two hours did not take much time.

C

So that is another piece of information you need to do that, and another thing is this: in the main layout: we are placing this access control, this razor component and in the nav menu. We are using this one and coming back to this appraiser. This is where we are saying that it was like this. It wasn't using the authentication. Now we are saying that authorized view and if it is not authorized, we'll say that you're you're not able to use it.

C

Okay, if something is not one we'll come in here and this one, um that's about uh the changes which you have to do see this the sorry north ordinance and then corresponding the access token, what we get it from the identity. We need to pass that to do a exchange with the api. We need to get the content from the api. So what we do is uh we need to register an api either. We can do that or we already have it isn't it.

C

So these are the two pieces of information which we use it and in the api we come and use this one small line in the startup.

C

This is one piece of line which we need to which we need to use it.

C

The work from home and the school is also from home, so you might use here some background noise. My son answering the question to a teacher to his teacher. I'm sorry so definitely you'll hear some background noises. So this is the one one piece of information we need to add it and and that's it and then you will add, use authorization use authentication. These are the two pieces after routing this is the order. It should be just before the end point.

C

So just before the end point, it should be routing and use authentication and authorization and add this stack, that's it so we go in here. um So you see that so now it is all configured now. We will come back to this one where we are adding the message analyst. So the message analyst will ensure that it is sending the um it is, sending the access token. Instead of the id token. Let's see that what is access token and what is id token, we will speak speak on that.

C

But there is an issue defining the default audience because at this point of time he mentioned it that what was happening is. It was not passing the proper proper id token and I was getting this four zero four zero one. So it was happening uh for a long time and he has given the work around also, but uh he told that the.

A

Good news is: this is sorry to interrupt quick time check, we're already around 12 15 now, assuming.

E

A

And then try to.

C

Take a few questions.

A

Before we wind up yeah sure.

C

Sure yeah, this is on this only one I just wanted to highlight. Okay, so let's go and use our application and see the difference between access, token and id token, and then I think we are good there. We can go on a a session.

C

I think we should make a half day event if I'm not wrong.

A

Yeah, you also need to defend what is a half.

C

So what we will do is whatever I wanted to highlight is not coming in here. I will go to incognito mode.

C

Now uss, so you see that whenever the laser basm you use it for the first time but which you are not seeing here, which you are not seeing here in this, the main one. So when this is the one which I'm accessing for the first time, so when I'm accessing for the first time is that it will bring in everything, see, I told, isn't it if you put it anything in the app settings.json, it is not at all secure.

C

If I come and click it that guy will get every piece of information whatever I am using my security codes, everything is out so actually technically. I should use it some c sharp code to bring in from an api some mechanism. Might we need to add it we'll look into that in future episode. But this is the security stuff do not put anything over here and every c-sharp dll will get downloaded to the, because this is a single page application.

C

Isn't it everything will get downloaded okay, including the uh blazervasm and net everything is coming in here. So I will clear this and again I do a refresh, and now you see that only the things which are needed, it's bringing in that those pieces but rest of the pieces. They are not. So I just click on this. It is not authorized and again I will click on login. So the moment I click on login. It goes and hits the auth 0. So I use this demo user one the same password I use universal.

C

So we do that and what happen is now it goes and authenticating and it is coming back to the login. It shows that welcome demo, one user and now I will go and speak about the tokens uh here- the couple of tokens here so so we have this id token and auth token jwt io.

C

So this is a access token.

C

I will drop it here.

D

Remove that and that.

C

Okay now see this, this access token always use access token to do an uh handshake with an api. So this shows that audience and it's it's a little bit secure. I mean some of the pieces, it's spitting it out, but that's what it is uh use, access, token and I'll take another instance of io and what happens if we use the id token I mean we can't protect ourselves, but, yes, we are seeing what best can be done in this agile world.

C

Okay, so now see that here you are getting the nickname of the user and the name and what are the credentials which we use?

C

A lot of information is coming out of this, so always when you do an handshake with an api use, the uh user access token, so you're not having those things here in this, the claims and other things will also come in in this piece, so always use id token at the front end or the place, the main app, which is using so just to understand that what menu options we need to enable do.

C

I need to enable the book list, or do I need to enable the fetch data or counter are those kind of things okay, this guy has logged in and what claims it is coming. Where should you use and all those things- and this access token is just you're, giving a proof to api, saying that hey, I'm a legal guy? I authorize myself. Please accept this access token and give me the content. Okay. That being said, let me come back over here and do a clear to the network and I go and uh come here.

C

uh It is bringing in this content for the first time and I go to the content, even though it's a c sharp see that, even though it's a c sharp it's not executing, if it is an asp.net, mvc uh web forms or whatever it is, when you click the button, it goes and executes the code on the server. So here see that there is no network activity happens and another important point is: I do the fetch again see this?

C

It is bringing in from the cache it's not making another call, and you see this book list when you click on the book list, it is going and doing a handshake with the api to bring the content command, don't fail.

E

C

Okay, thank god, did you fail? Okay, so see this is bringing in the content, and the other important point you can see is it is sending the access token. It is not sending the id token.

C

Sending the access token it's not sending the id token okay and another important point which I want to highlight here- is I'll. Take quick minute couple minutes to speak about the apim. uh So this is an api management, and how do you configure here so you'll go to this apis and you will have these funny facilities create a blank api or open api, spec or wsdl, your old soap or xml? I used to use in 2003 four five, six, seven and eight and wsdls and logic app service.

C

Azure functions, we'll see that in in a future episode where we are doing the functions in api management- and this is the things and here couple of important points. What we can do is I just want to highlight that I am here. I have an access to this api management and I'm heating one two, three, four, five, six, I write a script.

C

I can bring down this server so in that inbound policies, uh something you know you have this known as a rate limit, so you can say that within one minute from a particular ip, you can't hit more than five times or 10 times, based on the user's journey. How much, how much time you need to place an order in amazon or any e-commerce shopping. So, based on that, you need to decide that how many calls I can expect that from this particular user.

C

So you can use this one and you can have an ip filter and allow cross origins lot of other things are there, but please do explore this one and we have the jwt world uh validation. Also, let's go and quickly see that um so the jwt validation is saying that, if uh cross origin allow this from the website or the local and as I told we have opened it for a while just for the demo, please do not do this in your production and uh and another important jwbt validation is pretty simple.

C

All you need to do is uh go this, bring it this configuration and give the audience that's it now. It is going to take care of your access token. So if I come in here- and I don't send that access token- and I do a hit- it says that it's not. This is the message. I've given here, please pass r0 axis token: that's what I'm getting in here. Okay, the other mechanism is to do it over here. So it says that subscription key is missing.

C

Another important feature I would like to highlight over here is come back and upload the certificates.

A

Okay, I think vishwa we are already yeah.

C

A

Yeah, it will be good to bring it closer and maybe take a few questions right.

D

Yeah yeah, so I know it's a very vast topic.

A

Deployment options and security options, etcetera, etcetera, right, yeah,.

C

We will go deep.

A

On that, for sure yeah.

C

A

Okay, yeah, uh before, as we take the questions, I also have put a ticker in the bottom of the stream right, so please go ahead and fill in your feedback. If in case you think the content could be made concise or I know if you have any other content options. If you want to hear from us or any sessions, etc, please go ahead and pass on the feedback.

A

We will take it up. That's one part: that's the important part of taking the feedback, and the second part is in the beginning. Like I mentioned, we will be doing a raffle and then I will be picking a couple of winners, a couple of uh folks from there from the raffle and then they would be getting a just branch license, so you can download resharper or whatever the rider or whatever the tool of your choice. Okay, so please feel the so even for that I would need the need to know the identities of the people.

A

So that's one of the reasons why we are capturing the email address in there in the feedback all right, so please go ahead and do the feedback. As we uh answer a couple of questions here, okay, I see only.

E

A very few people.

A

Have given around seven to eight people- and I see them- there are more more people are attending, so it will be good if you all could go ahead and pass on your feedback. There.

C

Am I eligible for this license or not.

A

Yes, of course, okay, you should not be rating yourself in the feedback. Okay,.

F

A

F

A

Other question: let me just bring into the stream so yeah do we need to separate backpack for each database and deploy them separately. Here's another questions.

C

Oh actually, what happen is the backpack the project itself? You see this. The project itself will be one single project. uh This will be the this will be the project. Okay, while you are deploying it to the resource resource group, you will have the release pipeline, so each of those release pipe each of those release pipeline will be pointing to a re to a different resource group or a different subscription.

C

C

Once again, azure release pipeline so serious here what happen is when you are going and deploying it so the release pipeline will go and deploy to different environments, but the project, whatever the sql server project, is only one in one project. Okay, so you do it because the same set of tables, stored pressure, the content and everything should get deployed across multiple uh environments.

A

Okay, yeah, I think it's more around different databases if I'm not wrong or if you understand the question properly. uh So let's say: if we have multiple databases, should we create one backpack project per database.

A

It would be right. Environments are different anyways, we can do the promotion using release pipelines.

D

A

See if you're having two different sql databases right, can one dashboard project deploy into both of them or do we need uh two different backpack projects.

C

I think we have to go with two, because what is happening is here when we try to deploy. We are targeting to one particular database if you have two different databases: let's assume that the customer and professors like this is the e-commerce shop, and this is a student or a college shop. If you mix up both this in a single project when you deploy all the artifacts from this particular sql server project gets deployed in that particular database.

C

A

I think the target is at the connection string level, which would point to a specific sql server on the database. There right yeah and you know I think, that's what is that.

C

And an important point is: why do I need to deploy the college tables inside the e-commerce database.

A

C

About the bounded.

A

Context perspective: yes, it it validates that principle right. You need to separate things and then each microservice should have its own database. So technically you will have one backpack project catering to that particular micro service, related tables or so- and I just greatly mentioned. If you take an example of college and the other one. Yes, you don't want to do that way.

C

Right yeah, I have a separate database, I mean yeah.

B

C

Single responsibility.

A

Correct and another question is yeah: can we discuss transforming app setting or build configuration based on deployment environments.

C

App settings build content.

A

Yeah, this is what actually was pipelines in azure devops. You could do that right, you, you can define different enrollment specific variables and then you can change your app settings appropriately if you're, using azure, devops or if in case you're using github actions or anything else. Also, it is possible all right. uh Go ahead. Good I'll I'll I'll talk afterwards.

D

You can go ahead.

A

Yeah sure yeah what I was the thing I was going to tell this. Even with the code projects also, you can have different json files, it app settings.development.org app settings dot, whatever your production.json and different things. That's also another way to handle the environment specific variables, but if you want to maintain only one and then want to change it dynamically, yes, you can use azure, devops pipelines to do that.

C

Yeah, this is one of the way we used to handle this.

E

C

Tokens so uh whatever the json files are the config files you bring in all those config files, and you tell the you- will have this prefix token prefix and then based on the variables that will go and uh replace this one. You will have a variables here in the release pipeline or sorry bill pipeline in the ci bill, so we used to have the gated builds and the ci bill, so the gated bill gets triggered. When you raise someone raise the pull request.

C

It needs to know whether the code is in a compilable state and we run the sonar queue and if the number of issues are more than the quality gate check, it will break the build. And if the unit test is not more than eighty percent, it will break the build lot of bill breakers and we should ensure before doing the code review. That code is in a compilable state.

C

So in the ci bill is where we go and replace the tokens and do all those kind of uh things so that you know the artifact has got the the file so based on the environment, where you're doing or you can do that- one if it is a common. If it is a separate again, you need to handle it in the release pipeline. The as uh swaminathan was mentioning. We used to have a separate parameter file, as well as a json file for each environment, so that will be easy to handle.

A

Yeah another question from uh is not a question just to ask: is that can these documentation links be shared, and so what I would request, if you could add these reference documentation, links in your github readme right and then that would be good and we will share the github uh link with that with the audience I've already posted.

E

A

Of the github url here so if you could update.

C

A

With these reference documents, that will be helpful, sure.

C

Sure I'll I'll definitely do thanks. It's mine.

A

Okay and yeah, so this is more of a kind of feedback, a future session, I believe, can you expect azure, ad config and code walkthrough for web and api, app app service, environment and virtual networking with sample reference architecture? Yes, that's a very good uh uh thing to cover something they will try to create some sessions around. That.

C

Yeah, definitely I mean uh what we can showcase is we can take an application, angular application and and show how we can configure then in r0 how we can configure in azure ready how we can configure.

E

C

uh Identity- third third party idp, so they know the difference uh talking to multiple idps. We'll definitely do thanks.

A

Another question is how to prevent csr of attack in web api.

C

Okay, that itself.

C

E

A

There are a few annotations or settings which we can enable, I'm not getting that. On top of my mind,.

C

That should be here uh cross site, request forgery. It is not possible once again I'll. Tell you I'll tell you.

C

Auth board flow with pixie.

C

With pixie- and this is the one which is going to protect us from uh with this- is the one which is going to protect us from uh cross right uh cross site, uh request forgery. If I'm not wrong uh once again, for.

C

So recently, they've added this, this pixi flow and asking everyone to use this pixi flow, and this will protect from that cross, cross site, request forgery because they will add a piece of thing to this tokens, whatever it is being retrieved and that will get validated if it is coming from the other other ui.

C

I was just going through this documentation somewhere. I remember, thank god. I read it last week, so I was able to answer. If not, I would.

A

Have to switch off this is an interesting question. Is blazer good for pw apps compared to react.

A

C

I can take a stab on.

A

C

A

Is is an evolving technology now, right and and the kind of uh yes both are for building sps, which is single, page applications right, but both solves different purposes, and it's not like an apple typical kind of a comparison right. Blazer is not there to kill, angular react view or anything like that correct. If you are having a dotnet background, if you want to get into your, if you want to get into the front end kind of an application development, you don't want to.

E

Learn any other.

A

Programming, language, like javascript or typescript, or anything else, you could leverage your dotnet knowledge and then build with blazer, right and and since blizzard is a single page application. Definitely it also supports the pws, which is a progressive web, apps right, so it it really boils down to what kind of apps we are building right. I I like this statement, which one of my colleagues mentioned. It are we looking at building the fastest application in the world, or are we looking at building the application in a faster manner right?

A

That means, if you are a dotted background, you don't have a lot of learning curve. If you are building applications on blazer right, it really depends on what kind of functions you want to implement and then and and that's what would define, which would be faster or which would be better over there right.

C

Yeah yeah and another another thing I will I will speak is yeah technology point is one which you already covered swarming and another point is we need to do a technical feasibility study and performance testing of a p0 use case before choosing a or b you know I selected blazer and I'm seeing that blazer doesn't support a particular thing or react doesn't support a particular thing. Then there is no going back. You already committed your four sprints or five friends into the project and um and you'll be beaten to death.

C

If you say that I've selected a wrong technology, I need to go back and do it in another technology. Isn't it so just to be on a safer side? You need to do a take feasibility study before choosing it and and at least a performance testing of a p0 use case and that to very toughest uh use case in your entire project.

C

Do that technical feasibility and a performance study on those one or two toughest use cases, not a simple screen where you go and talk to a sql server flush it some table onto the screen. That will be easy. Anyone can do it so check on the features. What is needed and check whether this particular technology supports it or not. So that's very important.

C

I mean it's all like trial and error. You need to try and then you need to come. I mean I mean if we start thinking about that see in dotnet, lot of dlls are getting downloaded onto the browser. We saw that just now and in one of my tech talk the audience they asked it. Can we do a debug as we are doing it in angular, and he said that no blazer doesn't c-sharp doesn't surprise laser doesn't support.

C

I said that it's not the blazer, it's a google chrome who's, giving this tools to get into that javascript or a typescript. It's this tool. If tool they.

A

C

That you can read the c-sharp code of the dll and get it to do a f10 or f11.

A

They can so different things yeah and just to add an on the debugging perspective before dot net fight blazer was not having that particular feature, but with dotnet 5. The debugging experience is it's a lot better from visual studio or visual code itself. You can do an fa and then you can put debug points and then you can. You can debug a blazer or some applications running in edge or chrome or any browser for that matter.

A

E

A

Anyone, whoever has started working with blazer in dot net three, yes, that that time, the only option for debugging was right. Like console logs, you need to.

B

Write your blogs and then do it right.

A

Yeah so one last question we'll just take and then we'll wind it up. I know we are already beyond the time, so the question is: does it make sense to use active directory for authentication on top of odd zero, just to allow only employee block a particular department to make use of api.

C

It's like a you, have had a chai in a hotel and you're paying bill twice.

A

Yeah, you wouldn't want to do that actually see azure ad has it. I mean it's. It's by itself is an uh authentication authorization, provided you haven't built in an idp inside it right. So if your users are part of the ad, it wouldn't really make sense to create anything for with the not zero or any other third party things right.

A

If, if your, if your customers or your users are not part of the ad group and if they are somewhere else outside, then only you might need to look at any other third party idps to onboard them right. Let's say if you are creating a developer, focused website, so typically all of these websites, you would have seen sign in with github or signing with linkedin etc. Right. Those are some of the options, but yeah you would need to choose. Who are your audience and then what kind of roles you want to see and then how?

A

Where are they? They present? They? There are also options like azure b2c, where you can also onboard the customers, third-party customers into your ad etc, but yeah. It wouldn't really make sense to use two of them if we are not able to differentiate these different users or the user base.

C

Yeah see you, can you can do this one?

C

You can have a lot of identity providers, whether you want to use the uh sign in into this web app using uh azure id or this uh facebook or google or twitter like what we can do is go to this r0 and there they are asking you to sign up using the uh acme or uh continue with google. If I, if I go and click.

E

C

uh Login, so this guy will give you give us a set of options like we can use our own identity provider or the partner, sts security, token service or a partner id piece. uh Maybe the google or github or microsoft account, or this is the one which he is providing so this uh this is a different set of things like I can have my own user username password database, which I can validate. On top of that, I can go ahead and ask him if he is not a valid user in our database.

C

Ask him to go and get a token from any idp, so different things. Okay,.

D

C

I think I understood yeah. I think I understood raul's question. If something else is there rahul you can get back to us, maybe in the uh meetup or somewhere I'll uh definitely revert back to your query.

A

For sure sure, so I've got the list of people who have committed the feedback and I will be doing the raffle uh right now and then I'll share the share. The winners. It's going to be pretty straightforward, so we've got close to 16 folks, so I'll be using google random number generator to just pick up randomly a couple of folks and then I would start their names now, okay in the meanwhile, if you want to answer one another question, you could do that.

A

Maybe, since you are an anxious guy I'll post this, for you also okay, suggestion.

C

A

How to do app settings or configuration transformation for angular applications, unlike in dot net? There is no direct support now.

C

Yeah, so uh there are a couple of uh things um in the node.js in angular we use that env files to a decent extent and then in angular. Specifically, let me share my screen. Please flash.

C

It um in angular itself, under the vishy pylor um speaker, series.

C

Nice conference and v.net, I think, in year, 2020 we did easy, conf and then said the source angular web app here and says the source and the environment you can. You can use this to a decent extent, I'm not telling that this is safe, at least to a decent extent. You can use this like when I'm when it is a environment, it is the local thing and if it is hitting the prod instance, you can give the apm instance where you want to hit it and the corresponding this one.

C

This is one mechanism and another thing which I slightly touched upon is uh blazer. I told that the app settings whatever we are using is it is spitting all those uh secrets. So definitely it is uh keys of the kingdom. You are giving it so what we have to technically is we need to have an.

C

Api mission, so once once we start, we should go and call that api to get the configuration. So now the content, all the credentials and everything is in in memory and then configure your app based on that. I I technically I didn't do, but I've seen a couple of projects doing that, so there should be a I mean we did it in one of our atm application back in 2012 where, um where this uh asp.net waveform application, it goes on the boot up.

C

uh When you start in the 7 am in the morning when it starts, it goes and invokes the api to get the configuration like the language, what languages it has to support what payment methods it has to support. Do I need to support the check today, or only cash or only credit card, for all these configuration the start of the day configuration we should we will make an api call to bring it. So we need to do the similar thing in angular. I didn't do, but I have seen uh others doing it.

A

Okay, great, I think, we're almost towards the end of the session and I've got a couple of winners and I hope you guys have provided the right email ids in which, with which I can reach out to you and then share the licenses right. So the winners are ramesh, subramanian and prasad kumar.

A

All right so are the two guys who are going to get the uh license their branch license. So I will reach out to you over an email and then I will share the license code and the process to redeem these licenses.

A

Okay and rest of the others. Thank you so much for all your honest feedback. We will take it up and then we will plan our future sessions accordingly, right and yeah. I think that brings to the closure of the session. Thank you so much vishwanath for sharing all your knowledge, and I hope everyone enjoyed the sessions as much as we did in planning and then delivering it so yeah so much and we are available all of our github urls and everything was shared.

A

So if please feel free to follow us and github or twitter and then reach out to us and then we'd be happy to help with any further questions. Yeah. Definitely so.

C

We are waiting for the calendar to change next month and come up with another different architectural stuff.

A

Yeah all right, then, thank you so much. Everybody have a have a good weekend and have a great.

E

D

A

Stay safe all right, thank you. Bye.

E