Jakarta EE JakartaOne Livestream 2020, 4 Jan 2021

Previous Meeting

Next Meeting

⏯

youtube image

►

From YouTube: What We Learned from Porting PiggyMetrics from Spring Boot to MicroProfile | JakartaOne Livestream

Description

Emily Jiang & Ed Burns role play their experience into real world application

PiggyMetrics is a popular open source end-to-end sample which demonstrates the use of Spring Boot and Spring Cloud features in a microservices-style application. Spring Boot and MicroProfile are popular competing frameworks for building apps in the cloud-native microservices style. Functionally, architecturally, and historically they have many things in common. From a business, economic and governance perspective they have significant differences. This session from Java Champions Ed Burns and Emily Jiang, respectively of Microsoft and IBM, briefly surveys the history and non-technical aspects in comparing the Spring Boot and MicroProfile stacks and then will take you through a real world case study based on PiggyMetrics. We will share our experience of porting it from Spring to MicroProfile.

PiggyMetrics models a personal finance application and uses cloud native microservices features such as externalized configuration, aggregate logs, service metrics, security propagation, and distributed tracing. The porting exercise utilizes MicroProfile features such as Config, Metrics, Health Check, Fault Tolerance, Open Tracing and JWT Propagation along with Jakarta CDI, REST and JSON Binding.

Join Ed and Emily for a fun and informative compare and contrast ride.

Get Social!

Website: jakarta.ee
Subscribe: https://www.youtube.com/user/EclipseFdn
Follow us on Twitter: https://twitter.com/JakartaEE
Like us on Facebook: https://www.facebook.com/JakartaEE/
Join us on LinkedIn: https://www.linkedin.com/groups/13597
Tag us: #JakartaEE

A

And it seems we're alive again, so um uh I am super excited again to welcome um another, um a set of speakers this time around. uh We saw emily a little bit uh earlier today um and uh yeah. Emily was with us last year, but uh ed was with us uh uh last year as well. So uh having them uh in tandem uh in this session is, is like a bonus part.

A

So, as you know, uh ed is uh coming from uh microsoft and emily is coming from ibm and today uh they will talk about uh the micro profile. So please uh go ahead and take it over. Thank.

B

C

Happy to be here.

C

Let's take it away.

B

Yes, let's get a start.

C

Okay, so thank you for joining us today. My name is ed burns and I am very happy to be coming to jakarta. One live stream 2020.

C

uh for the past year and a half or so I have been working on java on azure at microsoft. Emily I want you to introduce yourself.

B

Yeah, hello, hello, uh I'm emily young from ibm, I'm the liberty, microservice architect and also the senior leader for microprofile.

C

C

So we know that you have, uh if you've been with us uh from the morning on, thank you for sticking with us on the stream. uh If you're coming to us in the middle of the day, then we are welcome and happy to have you, but we also know that uh there's a lot of other things. You could be spending your time on uh right now and so we're happy that you're spending your time with us in the next 45 minutes.

C

We would like to give you an introduction to to the peggy metric sample app and the port. The purpose of this session is to sort of share some lessons. We've learned in porting the peggy metrics sample app to run on microprofile from its origins as an example in spring boot.

C

So this is kind of a lessons learned and we're going to go over both sides of the story here with some demos and then we'll have a conclusion. So we hope to have a little time to take the questions.

C

So please, if you'd like now to click on the ask a question button at the bottom to the side of the number of people on the uh stream, and you can put your questions on there and I'm hoping that we will have someone to help us with uh curating those, and uh so let's go to the next slide.

C

Now we're going to try to do this in a more fun fashion, we're going to do a bit of role play here. So um in this role, I'm going to be acting as the corporate line of business app developer.

C

uh What I have done is inherited a spring boot-based line of business app and that's this micro profile piggy metrics app and what I need to do is to convert it over to microprofile and the goal is to reduce operating expenses in terms of how much infrastructure we need to run how much we have to pay to run it and also maybe hopefully lessen the cognitive tax that we have to pay on understanding lots of different uh apis from different sources and different vendors.

C

So that's my role. Emily's role is the micro purpose, microservices expert and also the microprofile consultant, and so she will be sort of explaining to me as the line of business app developer, hey here's what you need to do to move this thing over from one to the other.

B

Right so ada, and I will work together to look at the spring boot based piggy matrix to see whether we can improve it or not. So just give you some background on the piggy matrix, so piggy matrix was initiated by alexander unlucky.

B

If I got his name right, so it's a kind of a demonstrating micro service architecture pattern. So this article, the microservice architecture pattern was published by james lewis and martin fuller in the back in 2014.. So basically it's kind of tell people what kind of microservice architecture should be like.

B

So basically them like, in essence, is that I, a single application should compose of um consists of uh a few of um like a small services, is called microservices, and each of these services should have its own process and his his own release cycle, and these services will be communicated over like a rest, and each service should have its own business uh like capability uh and can be deployed um like independent, independently and also is a kind of um there.

B

Shouldn't be much centralized management, and the other thing is this: like macro services should be in, like should be a language uh agnostic. So basically, it can be developed by using like a different languages like java, go and python, etc. So you need to picking matrix.

B

It has like a stream like us services, at least three services. Actually, the four services um like a counter and also statistics and a notification, and each services has like its own database. This is very important, like each micro service should have its own database, so don't build like a multiple micro services share like a single database, so that is called like a distributed um like it's. It's called like um nightmare micro services. Eventually, you you, you just shoot yourself in the feet.

B

So if you want to check out more about picking metrics go to this um github repo to like a phantom move.

B

So this is another big architecture, uh including the like infrastructure. So basically the client directory uh like uh expose the endpoint and then then is the is the front end. There's a api gateway, so the client directly talk to api gateway under api gate. We talked to these three micro services and, as you can see, the micro services using ribbon historics and also it uses um like a config auth euro code and some other like dashboard and etc.

B

So so so this is the kind of the whole big architecture so add. So it's this is kind of the piggy matrix and for you to like um at first get it running. First.

C

Right, so that's a great thing for you to show me. uh You know, as a consultant, I'm happy to see that so what I'm gonna do is show you where we're at and what I've inherited to run speaking magic piggy metrics on spring boot. So let me just uh get that demo going. You know. Let me bring that up and we'll we'll get that going.

C

Okay. So what I have so far is uh it's a maven package. You know maven based application and I'm going to run it and deploy it on the azure spring cloud service. Now this is working really great for us, uh it brings all of the uh ease of management and um you know maintaining the stuff on the server. For me. I don't have to worry about any of the infrastructure, so I'm able to deploy this application to uh the azure spring cloud.

C

So I'm using the azure cli here and I'm setting up the different defaults and I'm creating the different microservice apps. You showed me there and then I'm deploying them. So I get the gateway, fat, jar spring boot jar and I'm able to quickly deploy that to the azure spring cloud service.

C

So this is going to upload to the azure portal, and so now I'm inside the azure portal- and I can see the fully managed platform as a service. I've got my three different, uh tiers or microservices running here, the gateway. Let's take a look at the gateway. You can see that it's starting to consume some resources now and I'm going to set up an external domain.

C

So I can access this gateway front end there. It is that nice, azure microservices.io and there it is. This is what the piggy metrics app looks like I'm going to go ahead and create an account and uh tips and tricks and set up my password and register, and uh you know going to be able to go in and log in and start to put some data in here.

C

So that's a very simple demo: you can see what we're doing it's an account management, software, personal finance, tracking software, and I was able to deploy it pretty quickly up to microsoft, azure. So let's go back, and maybe you can help me understand emily. How can I take this demo that uh runs really well and easily on azure spring cloud service and uh move it over and take advantage of the micro profile and jakarta e apis?

C

So after working with this a little bit I stood, you know I did recognize that spring boot uh is true to its word. It's a way to boot yourself into spring, um and when I looked at the competitive landscape of all the different um microservices technologies in java, it really is the oldest one. I mean I remember when spring boot was like a shoot off from spring roo spring room.

C

Most people don't remember that anymore, but uh it was a a sort of um thing that came around when ruby on rails was the hot stuff and uh it really kind of predated the whole microservices thing.

C

So the thing that I like about uh spring boot is that it gives you the entire spring framework, but on the other hand, that's uh quite a big thing uh and kind of antithetical to the microservices way of each system being light now, yeah, it boots up fast and that's all awesome, but the cognitive load is still there where all of the apis that are in spring are a part of the the mix.

C

So it's not as uh modular as say. Maybe uh open liberty would be, but anyhow it does work really well with spring cloud, and it follows the spring development governance model. So uh that's a great thing: it's a single vendor right, so you're you're, just using the spring apis. For the most part, you don't need to depend on any other jakarta or java x apis, but on the other hand, you with only one supplier. There is no separation from interface and implementation.

C

Spring is the implementation spring is the api. So we want something: that's a little more flexible and has a little less vendor lock-in. So we've heard about multi-cloud we want to be able to have our data estate be able to be portable across different clouds, but what we actually also like is having independence of the runtimes within those clouds as well. So it's multiple levels of independence here, so uh one vendor versus uh many vendors spring boot uh versus micro profile.

C

So why don't you tell me tell me a little bit more about what I'm, what I'm trying to understand here. Emily.

B

Yeah great uh yeah, thank you for sharing that uh kind of the concern, status and etc. So so, with that, uh like I highly recommend to seek some alternatives so like the spring uh boot is, is awesome, uh uh however, there's another alternative in the past, as you you mentioned earlier in the past like a spring boot, it was the only one like a cloud native programming model. Now, with micro profile, we can have choices, so we should seriously look into like a micro profile.

B

The one you're concerned is about one vendor and with macro profile like it's an open standard, there's many implementations for you to choose from so you don't have the worry about vendor, locking anymore, because you can freely move your micro service from one vendor to another one without uh changing your java classes and etc.

B

So let me like um give you some background introduction about this kind of comparison. Actually, last year, I spent some time to like um take a closer look between micro profile and spring boot.

B

I published article and under like eclipse news letter, so I highly recommend you take a look and read: read about it, so you are not a shell, so I give you like a quick run through about this, like a comparison so in micro profile, and we have a starter spring, which is comparable to spring starter door spring door. I o and start those micro provider io, give you a way to create your micro service.

B

So it's a the other thing for creating micro service is very like a martin folder earlier set is the kind of the most popular one. The rest um course so in micro profile, we have a jack stress to do your rest call in spring. You have a spring mvc and for the loosely coupled micro services, you use the dependency injection in micro profile. You have cdi in spring, you use spring roc or dependency injection for documenting your macro services in micro profile. We have a open api in spring.

B

You have a spring rest docks and for connecting micro services micro profile. We have a macro provider as a client in spring. You have a spring thin, client and spring mvc and for like um defining your like uh payload like a json json is very popular payload format so spring uh your micro profile will have a json b and adjacent json p directly from a decatur ee. In spring, you have a jackson or json b spring also support json b for reactive. This is for the like.

B

uh Up to now is a almost like a parity for programming and in a reactive you. uh We also have a comparable like technologies in micro profile. We have reactive streams operators. Also, you can use some other leverage like rx java and also we have reactive messaging in spring. You have a spring reactor and you can directly use the kafka api to inter interact with kafka and later on, for the like um handling hundreds of micro services, and you will need to do the configuration in micro profile.

B

You have a macro profile config in spring, you have a spring boot config cloud config for building resilient micro services in micro profile. You have four tolerance in spring. You can directly use netflix historics for security in microprofile.

B

You can use the jakarta yi security together with microprofile dwt propagation in spring, you use spring security and spring cloud security um and then for for the kind of intro interact with the cloud infrastructure like um use, a health like a micro profile, health um and also the metrics to add um kind of more application, metrics and also for them distributed tracing. We have a macro for open tracing in spring. You have a spring boot actuator for the health and matrix and for the distributed, treating you use a spring cloud. Sleuth.

C

Wow, this is a really great side-by-side comparison. It really does seem like everything I might need to do to port this application. There's an answer for it in microprofile. Can you take me through the code, and maybe you you know I've shared with you the code in advance? Maybe you can look at some of the code. I've had and show me how we can move from the spring side to the microprofile side.

B

Yeah greater question, so I don't want to throw you into the deeper end.

C

I'm swimming fast.

B

I also look at the piggy matrix spring boot version, so basically they have this kind of rest. um uh Like a course like a get and put you can directly convert to usage access and it uses all authentication, and then you can use the decatur e security and etc the ui. You can use the microprofile open api so actually.

C

Wait a second that means it works with json web token right out of the box.

B

Yes, exactly yeah, okay, so let me let her as I showed earlier. uh I also like um uh cloned that piggy magic, some github repo, so.

B

C

B

Yes, so this is the other services and etcetera. You have account service author service notification services, this statistics service- so you can see here. This is a source for file, so each uh like um service is have its own like a docker file, so it will create an okay image. So I I mentioned earlier like this is a thin client. So this is directly uh connected uh like like uh to do the client service architecture, and then you can declare like this thin client and then you eventually you can inject these thin clients.

B

So it's a you know. The similar pattern is used in all the macro services.

C

Okay, that makes sense.

B

Yeah, so let's go back to the slide, so after you I have looked at the code. Actually, I produced some like um initial protein for you to get so. As you can see I mentioned earlier, you use a thin client, so you can directly convert to use the micro profile. Rest client, like a registered razor client, directly correspond to that. So.

C

Those beans are just cdi beans right.

B

C

B

Beans, so is that the other thing is that in open uh liberty we have a guide so to get you started, so each guide is taking about a half an hour. So that is like a rest, a client. So I can directly show you like a quickly show you. This guide.

C

So these guides are designed to be easy for my team to take a look at, and you know, they're able to put this and get on the ground running right.

B

Yeah, exactly if you want to learn like uh jack stress, you check out this guide. Okay, you don't have to put a snippet, etc and then you can get a clone into it. So.

C

No, by the way, are these guides going to be updated to use the new, jakarta namespace.

B

Yes is the great question: yes, we we try to keep this guide up to date is a new like a micro profile, update and micro profiles back. We also updated this guide as well yeah, so they are not. They are always recent.

B

Okay, let's go back here.

B

So as I I can see as I mentioned earlier, so you declare uh that as the red is the rest client so directly, you want that interface. So this is the directory, the backend interface, and here you can directly do the injection, so you can in spring you use auto wired in micro profile. You use the art, inject resident client, and then you can call the back end directly from there.

B

So for the configuration- and you can see here that some must bring annotation at configuration. So you use the configure service and also is a is a macro profile. Has the corresponding, um like a technology called micro profile config, so you can directly use, add inject, add configure property. You can't define your property in any kind of configure source such as like microprofile dash, configurator properties or the system of variables or environmental system properties or environmental variables or any other like config map or the zookeeper and et cetera.

B

You know do anywhere again, like you just write, I put some like, uh like a uh guides, openability guides for you to help you. If you want uh like a learner, thank you uh and for secure endpoint, so is one service using a secure endpoint enable resource service um alternative is to use um uh the currently user security rules allowed to secure.

C

Wow, that's a very old api. There that's great to see that one still working.

B

Yeah, it's still working and functioning well and the yin micro profile will have provide a way to pass in these rules. We call it a microprofile jwt. So basically, it's um add the upn and also groups over and above the jwt token. So when the backhand receives them like a request, you will get a header or authentication header. So if you like open up that token, you can find out like the back-end runtime will be able to find out which group you belong to.

B

If you belong to user or admin, you are allowed to access this one. Otherwise it will directly seem forbidden.

C

Okay, so that's kind of like I'm able to access the claims. In the token that way,.

B

Yes, yeah great so, and you can also inject the token injected claim to the backhand as well, so you can see it so. I mentioned earlier like in spring also spring, like a support. Js research, you know and like that directly the dsr, uh however auto wired is kind of more popular in spring uh application. So, however, is when you use a micro profile, remember to use inject.

C

B

So this is a data access, so in this um pigmetrics it use mongodb. As you know, mongodb use is a nosql right. It's, um uh although it's uh you know it's currently, there's um a jakarta! Yes back is called janoshiko is against uh you, you are able to. You, can either use the java, no generical implementation to directly work with mongodb or you can use some other, like libraries and etc in spring. It has some like uh yeah apis for you to work.

C

With it looks like uh there's some common apis between these two, it looks like they both are using bean validation. Is that right.

B

Yeah, you are absolutely right. So acquisition supreme portal also supports some of the dakar ee specifications, as we mentioned earlier, like a gsr330 at inject, also the beam validation and also the subject. So, basically, if you think, okay, you don't need to migrate from java x to decatur x for the spring up actually you're wrong, so use you to migrate from my java x to d carter, if you're using beam validation or uh savage apis, um yeah, okay or they inject.

C

Let's keep going because I really I want to get started, I'm very eager to get going with the coding.

A

Yeah- and there is there is a quick question: if I may, uh is it possible to inject dependency by constructor, not by um inject in fields.

B

Yes, cdi supports three levels that, like uh injection uh one, is the kind of the field injection. The other one is the constructor injection and also the method injection. Yes,.

C

You mean I can still use this at post construct and at pre-destroy annotations.

B

Yeah you can use the poster construct and redestroy as well yeah. So that's a really really great question. Okay, so that's is kind of pretty much for your application. So ed is because you are using micro profile. We also have a really good provide a really good ecosystem with the cloud infrastructure. That's.

C

Really important because I want to be able to use industry standard tools like grafana and prometheus and zipkin and elk to understand how my application is performing and to be able to do diagnostic and log in analytics.

B

That's a great uh great uh thing, great comments at so with that you, when you mention graffano and prometheus, and you I highly recommend to check out these metrics because, like this magic produce the format and directly to be used by graffana or uh yeah permissions, and also is um there's a lot of like has some dashboard already on the grafana for for like open liberty, and some other vendors also has some like uh yeah dashboard, provided it's a it's a great it's a great uh like ecosystem, so check it out. It's very easy.

B

You can see here the annotations and etc.

C

Let's keep going because I really want to get to the code, so maybe we can speed things up.

B

Yeah, so the next thing is remember: you have these four macro services and then they will be uh running independently in the cloud remember uh like, in order to talk um communicate with the cloud, each micro service should be intelligent, they should be able to communicate with the health status with um like kubernetes.

B

So with that, I highly recommend to you you to use uh like a liveness and a readiness annotation to direct to say that and what kind of circumstances these micro services you are like.

C

B

So it's a wide example where you can say: okay, I met around around probably so this one has to be a like a down so that kubernetes can kill it. So the readiness you yeah you can you can like one of the example.

B

If you check the database or check the back end to see whether you can connect to them yeah, to see whether you are you're ready to serve a request.

C

Okay, well, that really makes a lot of sense. You're, basically saying I'm telling the container environment how to tell if I'm healthy or not.

B

Yes, exactly so this make a micro service is very intelligent, so the next thing is, you might consider to improve the resilience status of your micro service. As you can see in like the picking matrix the spring boot version you use, they use the histrix and ribbon like, like a have. You heard all historics yet.

C

I have, but I thought it might have been going out of service- is that right.

B

Yeah, you are absolutely right, so so with this yeah for the picky matrix. The first thing is that you need to worry about you for this kind of like a cvs and you you need to worry about yourself. You need to fix that kind of the the vulnerability to yourself.

C

Yeah, it's very important. I keep a very tight eye on the cve database.

B

Yes, exactly so you you need to take like a care of yourself, because this is a go out of the maintenance is, is a under maintenance?

B

Is no active development happening, so uh good news is you can direct and migrate this one to micro profile for tolerance, so info tolerance is kind of it has a retry, a circuit, breaker timeout, a full back and a book head, basically that all the functionalities offered by histrix is provided by microprofile, uh like a fault tolerance.

B

The beauty with this is that you only need to worry. You only need to call this api. The implementation will need to worry about the the cv and etc like the open liberty. If there's a vulnerability thing uh fund, open liberty need to fix asap.

B

So this is the kind of a beauty of this you you can like uh just need to worry about your own business logic.

B

So next thing is that the api, so basically it's the kind of the uh in spring you have in the spring boot you use api gateway, your micro profile, there's a micro open api. If you, you said you're, you want to like comment to use um uh open liberty openly. Pre-T has kind of uh like a swagger ui integration. It's a micro profile provider.

C

Oh, you just said the magic word. All my developers are saying: oh everything has to be in swagger, so that's going to work for me great.

B

Another take yeah so for the distributed treating you mentioned earlier in spring. You have automatically been done by spring cloud sleuth and similarly in micro profile, uh we have open tracing automatically managed for you, so you don't need to do anything, and this can work with either a yoga or zip key. So we have which one which one is your favorite.

C

Well, I tend to go with zipkin, because I have a developer on my team. Who knows it very well. He just came from another company we're able to hire people off the street. That knows zipkin. It seems everyone knows zipkin.

B

Okay, that's great, so this is a guide for this. Lip kit is a guide for yoga, so so one last thing for the infrastructure.

B

So is that, like you have uh you know in the pigmetrics you heard of this monitor dashboard and service discovery, so you can directly like a migration to graphano or kubernetes, zip king yaga, like the yellow, key and etc. So this is the make your application really clean.

C

When I go to sites like glassdoor.com or um stack overflow, I see tons of questions being asked and answered on all of these technologies on the right here. So it's really great that um my app will be able to take advantage of these things that everyone already knows.

B

Yeah, absolutely so you just like your developer, just concentrate on the business logic. All the others is pretty much standard cool, okay, so.

A

Again, there is a question from werner this time: what about micrometer can that be directly migrated from spring boot to microprofile.

B

Yeah, that's a great question, so it's a micro profile at the moment is trying to is kind of trying to investigate uh to support uh yeah micrometer so with hopefully next year. We can work on that. So with that support, that'll be a great uh yeah story to migrate from a micrometer spring boot to a macro profile. Micrometer.

A

Yeah so another one, have you tried to convert the spring integration test with a spring boot test? Do you have any porting tips.

B

Well, that's a great question for the testing, your micro profile. We like uh recommend, like uh some technologies technologies uh like a mock framework and also the uh the container testing uh like a tester containers, uh yeah.

C

Oh wow, so test containers can work too. Okay, yeah.

B

So check out the the other github ripple uh is being created for testing macro services is called a macro shed testing.

C

Yeah, well, it's good to know with only one sprint to do before we can get this work done. I don't think we'll have time to focus testing too much, but let me see what we can do so one sprint later.

C

Okay, I'm back, we built a prototype and it's just a proof of concept only with showing the different minimum viable product user scenarios. So we got this thing running on azure kubernetes service and what we did was leverage the built-in discovery from kubernetes. We have a sort of a trick that we've done to do that. You can take a look at our github for that.

C

So, let's go on to the next one. In fact, I'd like to give you a little demo.

C

Let's take a look at what it takes to run: piggy metrics on microprofile on the azure kubernetes service, we'll start by looking at the azure container registry, that's deployed alongside the azure kubernetes service cluster.

C

The container registry in our case has four different docker images: let's take a look at the account service, docker image, and you can see that we have the tag name and a convenient uh docker pull command.

C

If we wanted to pull it ourselves, we could copy this to the clipboard and just run it and let's go back up to the resource group and take a look at the test cluster and look at the workloads, and you can see, we've got all of those services and more because we have grafana and zipkin and mongodb, and we have the kibana running and the elastic search prometheus. All of these are different docker containers, but let's take a look at the ones that we have here: authorization statistics account and gateway.

C

Those are the ones that comprise our actual app I'll. Also mention that, though the azure spring cloud service has the monitoring you can get the same level of monitoring and information at a higher level for the azure kubernetes service directly here. But uh this demo is not going to be showing that, because I'm showing the use of zipkin, grafana and kibana for these purposes.

C

Let's look at the command line for the kubernetes system. We will show first the pods that we have running.

C

We have two each for the workloads that comprise the microservice, and then we have our deployment showing the same kind of system that we have actually deployments that are set up and let's now take a look at the.

C

Service and the service gives us the external ips for the things that we care about the first one we care about is the app gateway, the application gateway. Let's take a look at piggy metrics, so, let's log in with our test account here.

C

And we see our income and our expenses.

C

Let's go ahead and create another expense of type 10 a month on.

C

Transportation.

C

And click save we're just putting some data in the system, so we can take a look at it in the next phase, which is the zipkin, so we will run the query and see a few seconds ago we have a put and let's look through the span, and so we can see that we hit the gateway and then went to the account service and then went to the statistics service.

C

So if we click show, you can see this trace flowing through the system. Here, it's the jax rs component using the put operation, and you can see the different urls that are going through.

C

Let's now look at grafana, which gives us the performance information. First, we will take a look at the cpu, so you can see the system load being shown here. Another very interesting one for us and this micro profile world is the servlets. So we can see the different instances of the servlets. As emily mentioned, we have the health checks and the metrics. So there's a lot of services that are running uh servlets that are running here, to provide the micro profile introspection to do the work that we're demoing.

C

You can see average response times per request and over the last 10 minutes by the servlets.

C

Finally, we will look at garbage collection and, if we scroll down here, we can see uh you know we have not been using the application. Then we started using the application and do some more time per garbage collection cycle.

C

Okay. Finally, we will look at kibana and show our logs. So let's refresh the query- and you can see- we have a lot of data- that's been over the day that we've uh days that we've been interacting with the application.

C

So, let's drill down into this particular message here, you can see a new data point, so this is when we put the new expense into the system, and so all of the interactions with the system are logged directly with the elk system that we have here.

C

Let's go back to the slides and I'll take through it and we have the github repo. So what we've done is the technologies we've used. Are the config service fault tolerance, uh the java json web toolkit? We did the rest client. We did all the recommendations that you showed us and I will um go to the next slide, and so we've made the side-by-side comparison kind of showing uh what was on the uh azure spring cloud version with spring boot and then the corresponding um micro profile technologies.

C

And what happened is we didn't have time to get through the notification service, so we did gateway account, auth and statistics, but not the notification service. So, let's go to the next slide and so the takeaway. We really feel like um the sort of principle that we're trying to get across is that developers should be free to choose what works for them right. So enterprises should provide developers with the tools and platforms that they need that enable innovation.

C

So this really fits in very well with the micro profile way or micro services way of doing things. Each microservice can build their stack. The way they want to and both spring and micro profile can provide facilities for developers to build next generation applications for the cloud.

C

So there are some similarities and differences, but we're really happy to see that it's a lightweight run time. So, let's go up to the next slide and uh here's another set of uh comparisons that we've sort of distilled it down even further.

C

So in the microprofile world you have a thin wire and in the spring implementation world you have fat, jars or small jars and they can use the jib plugin for creating containers out of those things.

C

Let's go to the next one, and this is the one that hilmer was asking about. So the uh github repo for this translated thing is right here. um My colleague joan guoma is the one who actually did this and um you can take a look at that. The original piggy metrics on azure spring cloud there's an end-to-end example that really shows how to bring how to do that.

C

Whole thing that I showed two minutes at the beginning of the demo: there's a link to emily's migration guide right there and uh I'm sorry emily's comparison guide and then the two different starters that we talked about. One of the things that's really great, I'm stepping out of the role here um role play is to say um we're all learning from each other in the java community. So I will give props to the spring world for creating.

C

You know spring roux and then eventually spring boot and then creating the concept of a spring starter and uh having that idea that it's a symbol website you can go to and get your project jump started, and all of that uh that's operational excellence that spring really has brought to the table. But micro profile is doing the same thing. There's no uh there's no monopoly on innovation and uh what all of us are about spring uh microsoft, oracle, ibm, red hat.

C

All of us are trying to empower developers uh to do the thing that they need to do, which is create business value for their users. So, um that's why it's really important to recognize that, though we are sort of having friendly competition here, we're all working together on the behalf of developers.

C

So, thank you very much um we're gonna uh I'll work with uh tanya and see if I can get that video published that we had a problem with.

A

And also just to squeeze in uh last question here: what about the future of open tracing and micro profile over tracing and open census, recently merged to form open telemetry will open telemetry, ultimately replace open tracing.

B

Yeah, that's a great question, so we discussed this uh like a microprofile open chasing this year and we are prototyping to adopt um open telemetry uh like uh in the next year. Probably uh we will do a do more in anger, so, basically, we directed that can also directly yeah consume the open, telemetry uh and into macro profile. The other thing is because the open telemetry has a matrix uh also has the open tracing. So we also need to look at like the relationships between the macro profile metrics with the open, telemetry matrix as well.

B

So, however, we want to get started to consume the tracing first from the open telemetry. So what is this? If you want to, please feel free to contribute as well, so it's directly logan go to the logo issues and directly chat on the data of the individual individual macro profile. Specifications.

A

Okay, so um I guess what's left for me to do- is uh to thank you both. I will invite you to look at some pretty nice comments uh about this session uh uh in the chat. um I am super thrilled and excited uh and uh well not just comments about the session.

A

I would say uh some looks were commented here as well, so uh uh uh ed and emily people uh love the the the look so uh keep it up and uh yes, yes, we do expect you next year as well with uh so much more innovation and uh involvement from uh both both of you and both of the companies microprofile, um sorry, microsoft and uh uh ibm as well. I.

C

I can tell you: we have some very exciting things that we'll be announcing shortly.

A

Okay, we're looking forward to that. Okay! Well, uh thank you! So much and uh we will be in touch as always, uh and I will work on on uh making sure.

C

You're requesting, let's make sure that video.

B

C

A

Yeah. Thank you. Thank you. So much see you soon bye to both.

B

B

B