AsyncAPI Community, 26 May 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: AsyncAPI SIG meeting 24 (May 26, 2020)

Description

This is the recording for the AsyncAPI Special Interest Group (SIG) meeting #23.

Attendees:
- Collin Sullivan
- Fran Mendez
- Jonas Lagoni
- Laurent Broudoux
- Łukasz Gornicki
- Semen Tenischev
- Yacine Kheddache

Moderation:
- Fran Mendez
- Łukasz Gornicki

Agenda & Notes:
https://github.com/asyncapi/asyncapi/issues/384

Chat Dump:
00:48:01 Fran Méndez: https://github.com/asyncapi/extensions-catalog
00:58:27 Colin Sullivan: This is a very cool project - anything that shortens time to value in k8s is high value!
01:01:12 Laurent Broudoux: Thanks a lot Collin !

A

Okay, it's recording hello, everyone I welcome to us an API community meeting, I'm Akash your host and Fran is no longer a host, but he's still helping a lot. Today we have three items on the agenda.

A

We usually leave our items till the end, so I see that we already have guys from micro rocks. So please accept my apologies, but we will go through to those two topics that we already have on the agenda because I think they pretty quick and the the generator one is pretty important to handle um and I'm, not sure how long you're gonna do the demo and how the discussion will evolve. So, let's keep you for the for the end.

A

And just before, I start so we have Lauren and excuse me if I pronounced wrong. It's yes in that's perfect, okay, awesome, but so I know your guys from micro rocks. But before we start, we have also Colin Sullivan just to ask you if, if you're, together or Colleen you're from different reason on the call, you have some important question you prefer to ask before we start no.

B

No questions just quickly introduce myself. My name is Colin Sullivan I'm product manager in the Nats project, we've seen a lot of people that are really interested in Nats and async API, so I just figured I'd join the call run a little bit more about async, API and kind of you know just listen in for a while. If that's all right, yeah.

A

That's totally fine awesome thanks for joining all right. Thank you, um okay, so yeah agenda. So today we need to quickly discuss the topic of the of the generator asking API generator. So it's a just for for you guys. If you don't know this project, it's a it's a library, that's powered by a many features that allow us to generate documentation code.

A

So it's a I would say a kind of framework for generating code. So you basically feed it with the a sync API document, and then you can create templates where you specify define what kind of files should be generated if it should be HTML or or some code files. You just template them and there's there's a bunch of features that come with generator and it's already power powering our ascend API help the playground.

A

Many people use it and we are now in the stage of not yet major released over the last few weeks, it's pretty unstable and introducing the breaking changes all the time because we're yeah we want to introduce. Finally, the major release and the question is: what do you think so I when I was thinking about this? Introducing this release, maybe before the proposal?

A

Let me just show you so the current situation is, there are only four issues: the conditional files back I'm already investigating for the readme there's already poor request accepted by phone I'm, just waiting today for mark. If he has some more questions and comments about it, dummy like housing, API document, that's already in progress, I already suggested- suggest its solution.

A

So only this one is not yet started. So, as you can see, we are like almost almost at the end.

A

So that's why I brought this topic, so I would suggest that we we don't release like one zero immediately, but rather a release candidate and then, when we will do the release candidate, we'll just announce it share with the community, so everyone can can try it out, I make sure everything works and we would also, for example, we have our github action where we use generator as a library and the same with the generator we have in the US and API hub.

A

So I think on our site would be to make sure that the release candidate is used in the github action and the Assange API hub, and so we can also confirm that as a library, everything works as expected and the features are are not failing. So that would be my suggestion. Do you see any any issues here and I would suggest that the release candidate is, if not at the end of this week, then for sure they earned the next week. I would create it famous last words right.

C

Famous last words, when you say it's gonna be next week.

A

No, no, it's for sure and that's what next week.

C

So yeah I think it's to me. It's it's the right way to proceed to have at least candidate first because yeah by the way, sorry for all these breaking changes that we've been doing on the generator during the last month. But you know it wasn't even the worst one yet and we had to eat right fast to reach to that point right where we have a stable version of the generator and and yeah for those who don't know so much about the directories.

C

We call them templates like like Lucas a but but the templates are code, generators and and documentation generators, and it can actually be anything else, but mostly people use it for code and documentation generation. So and when I say people live in our selves right so so yeah and what I'm not sure and I've been thinking a lot about it is sold.

C

Can you can you go to the list of issues again yeah, the first one? Maybe we should probably leave this one out. I implement the way to generate classes or types from distant schema, technicians.

C

It really depends, I mean the thing is the thing is that.

C

We can, if we can manage to make this this feature to add this feature in the in the future, without creating a breaking change. If we assure that we can make it, then we don't have to include it as part of version one to make it stable, but but yeah it will be at least required to investigative if it's gonna be a breaking change or not right. So.

A

So, the last time we talked about this issue, we said like that we will for sure do some like deeper investigation, how that could be done just to make sure that it's not a breaking change in future. So when I'm looking at this issue, I, don't necessarily assume that I will implement like full-blown feature. It's rather like spent at least like two full days on research.

A

How this could be done and and see, because I have different thoughts like even I, was thinking about hoops, maybe so it's um it's just requires some investigation before we can actually say if it's breaking change or not yeah.

C

Exactly so it's just if it's not a breaking change, we can postpone it and add it as in the future near future. For sure we need to make sure that we have that that feature at okay, it's just so we make generator stable as soon as possible, and people don't see many breaking changes very often right. So so the library becomes more reliable right, not just it's just.

D

Because of that.

C

Otherwise, we'll have to be releasing v2 v3 before right very quickly, and that's that's okay with yeah, it's it's gonna, be a mess cool man, so yeah from my side, I, don't know and say man what they have to say, but they are the ones working the most on the generator. So, but for my side, disease, along with you.

A

So you know Simon anything against the plan for this candidate. One.

D

No and about conditional files I think it's better to consider whenever we need this feature currently or replace it by hook like in spring template, because currently it's too complicated to use.

A

Yeah so I'm I'm working on this book and I'm not really familiar with the of the future too much so I'm, not sure. Yet if it's complicated to use or not so let's see how it goes and if it's really not really handy because at the end, I know that you have this hook in Java template but I mean, wouldn't it be nice if you could, just by configuration, specify which files should not be generated. Instead of writing custom hook.

D

Yeah, my yeah, it will be easy in configuration, but then it should be more easier razor. Its now, because now.

B

D

Spend a lot of time to understand about JSON validation schema to use it.

A

Okay, then, let's see I just started the investigation, so that's how it goes and we continue sinking on slack. Maybe what do the next three.

E

Yeah, just a quick note on the json schema class generation, I'm kind of I'm kind of waiting for it, because I don't want to put too much effort into fixing the bugs that I have now with generating schemas and.

E

But yeah I'm a little bit of Swampert school as well, so it doesn't matter if it's like in a month or how long it's gonna take, but yeah um I'm kind of waiting on that for the Nets template.

A

B

And if there's anything we can do to help with the Nets template. You know certainly feel free to reach out we're happy hope over weekend. Yeah.

E

I, remember I also have some people that all already volunteer to test the pull request before merging the initial version is at least but yeah thanks our member.

A

Okay, awesome, so I will for sure consult with you Yanis when I will yeah work on this. This feature, okay, so d, so we are going with the bridge candidate release.

A

Next, one quick, if you remember I, asked why some slack, because on general channel we have like over 500 people, which is pretty pretty a lot, but I made two tests asking people like if they follow all discussions or rather check from time to time or just have the channel muted.

A

Roughly 15 people have like half real-time, seeing our messages have just checking from time to time and I was thinking like do you think, especially like you, Yanis and and Simon, because you you're quite part of the community. So do you think if there's anything, we should start thinking about changing in the in the in the setup of this slack workspace, like maybe granulate channels or like we have, for example, with generator, maybe with other topics or or simply just leave it as it is.

A

What's your opinion here.

D

I think we have not a lot of discussions in general or generator channel, so it's pretty easy to read it each day and yeah since not a lot of topic. It's for now doesn't require more more details or more granule separation in topics. From my point of view,.

E

So they all agree, I'm, not sure the capability of slag I haven't hosted a server yet, but eventually I would assume that we maybe need a template channel to separate to generate our question from the how to make templates.

E

um But yeah, that's it. That's kind of the only thing I have I, also like the setup.

C

I'm afraid that if we, if we divide things in topics much more, then things will lose visibility right. So the cool thing about general channel is that everyone is subscribed to general channel right. You have to cannot unsubscribe from general I guess so and we don't have too much noise there there yet sometimes, yes, sometimes in a day you might see like I, don't know 10 or 20 messages, but some other days you only see one or two messages. So it's like to me it's still pretty decent but yeah.

C

Regarding your questions and the fact that you only got 15 replies or that's funny, because it means that people are actually not reading the channel right, but yeah I mean I think that the issue is actually there like. What can we do to engage more the community right? It's not so much I, guess about how the channel knows about the workspace, how the workspace is organized, but maybe the content we push there and the way people can subscribe to that content.

C

Because, right now we have the newsletter, but some people might not be interested in newsletter as well. It's too much content will be great if we can have maybe on the blog or in the documentation, some kind of add-on subscriptions since, like this, some people can subscribe to me the connotations of new blog posts or new or new.

C

New guides on how to do specific things that people might be interested in more granular subscription controls right, like imagine that you join slack and you could have a button saying- don't notify me unless this they're discussing about this topic. But of course this is like kind of magic, because it's it's a free form of chat but but yeah and not on slack, but maybe another other things that we can offer like.

C

The blog like the website, the documentation or now like, even that, the newsletter that you might want to subscribe to specific topics only I know that, for instance, in you, the block I'm I'm following in 4q and I, usually receive the newsletter only for software architecture, topics I'm, not interested in security or in testing or in any other stuff. So I don't receive. I only receive like a summary of what happened last week on the blog and all the blog posts related to software architecture.

C

So maybe this could be another way that we can engage with the community, so they actually get what they need and yeah. But I think it will be great to ask them to ask everyone on the slack channel using type form or something so we sent. We send everyone, an email and with with the type form survey and.

C

We can, we can celebrate them and see because maybe some people just join the channel, because some day they were interested and they're not interested anymore, and they are not. They don't even have the channel opening their desktop, app or web app or anything they just not sign it. So they don't they don't even remember they. They have an accountant and they can keep a workspace. I know it was. It happens to me very frequently, yeah.

A

Okay, so let's the best will be first to ask questions to those two that that are muted. What's the reason, yeah.

C

A

As you said like me as well, when I joined a few communities to just engage first second or check from time to time, I just meet, and it's perfectly fine just to make sure that we're it's not because we're or screwing something with thee. The way I.

C

Think it's also because some people- and it's also the way I was using this like channel at the beginning. I- was using a slack channel to give support to people who started to use it in API.

B

C

Some people, many many people, just joined to ask some questions at certain point in time, and then they just quick. They sign out of the channel and after work space, and they don't see anything anymore so and because their intentions were just support, not really become part of an engaged community or anything right just solve a specific problem at some point in time and then just continue with your stuff.

C

So yeah will be great to know how many of those are actually this scale. If this kind of this kind of users would be great.

A

Okay, so the outcome I'll just plant or something of tea, how to communicate with them to get some answers possible, and so that's it and the next one is yeah so guys you would probably like to do some demo, so you will have to have host right and which one of you eats it.

F

Maybe I'm gonna be the host and let Yasin have some words to twin producers. Yeah.

G

First of all, guys thanks for bein, very glad to be am proud to be to be with you guys, because I mean what you you're building is pretty amazing and we, huh let's say very, very happy about the new features we are walking on and on both side.

G

Just maybe to give you a little bit of background regarding us and my crocks as a project, so this project has been starting started in 2015, and so we are working on it for quite a while, but really on a pure, let's say open source way which mean and before you guys ask.

G

We are both walking for reddit, but this project is not related to, rather at all, that's the pure open-source project and it's something that which let's say but part of a passion today, and it's just walking day on night on on it, and especially during the weekend. So that's it, and so it's my croc is a pure open-source project based on communities. I mean it's Cuban. It is native fully dedicated to only one one specific point to it, which is an API mocking and testing.

G

That's that's really what we are focusing on, because we think that it's a big pain for many users and especially for big companies we, which we have been able to to chat for the latest decade and our main goal is to turn API contract collections of UI and soon ackc API into mock in a second and ultimate goal is really to speed up API development and secure or lifecycle management.

G

So that's that's what we are scene on so now, I think and what we are really interested in is just to let's say to less to to do a short demo. The live demo of what we I mean the way we would like to support a CK API within my crocks and grab some feedback from you guys it's something which is part of a road map and backlog and more input and ideas.

G

You know request up whatever you come from, you guys will be very helpful so that we can integrate that for next release, which should be acing, API compatible and just from a usage point of view. We are not only doing that for fun.

G

This is also coming from a lot of users and integrator. We are working with who are already using microbes and would like to use async API, for you know, mocking big IOT project or project which are using a huge amount of of streaming requests and so on.

F

Sure one of the first customer of my clocks is indeed my former employee employer before joining reddit I was working for a financial services firm that was deploying a huge SOA program at that time. We do not speak yet about API, but we speak about web services and that's where I got this idea about having a central platform, a central repository that was able to generate mock on demand and incidence to allow the scale up to hundreds, millions of services or API throughout the organization.

F

So, as Jason said, the goal of my Crocs is not to define a new specification. New format is to reuse existing stuffs, so we started with so Qi project. Then we had it postman collections, then open API files and the ID is to turn this artifact. To turns the example and the schema information into this artifact in to life marks.

F

That is the first part, and the second part of the tool is about the testing, because once we are the these artifacts with all the schemas of the examples, we can reuse everything all these stuffs to test implementation of the API and to ensure that they are compliant with the specification. So that is the second part of my crocks, all the tests and the compliance ready part. Ok.

F

So, as you may have seen, we already support in my crafts, some some different kind of API. So, for example, in my repository, if you can see my screen, you may see that we we have already loaded some REST API. We also have a some kind of a dynamic API. That is a some kind of back end as a service API, and we can also able to to load soap web services as well and the next release.

F

We have a specific focus on a syncope I, so the idea is just to let the user compose described his API using your tool using, for example, the playground. Ok, so here I have a specified.

F

The simple API, that is, the user, signed up a PIR that consists in publishing an event, for example, and of course you know you are able to define some schemas for for the payload and the specific part about my quacks is taking care about examples about the real-life samples that you may put into the specification as a companion to your schema information, so that you've got real-life examples on how the API is expected to work so right here, I have defined two example: Rohan and John, and you you just see that, and that is the reason of the pull request.

F

I have to open some weeks ago. I have used this formalism. You can specify the headers of the sample and the payload so just used here, JSON format or Yammer format. Okay, and within these samples, you can also use some specific expression, language that is supported by my cock's. That allows, for example, the the generation of a random string, all the generation of a timestamp using the the now notation. So this is a notation, but it is already available for all the different kind of knockings.

F

So you may reuse this and embed this expression within your subway project, postman collection or the sync api. So just download the result, and when going back to my crocks, you can import the results. So there is different kind, a different mean of importing stuffs in my crocks. You can just same my crocks. You have to listen to this git repository, for example, and it will refresh the mock definitions, the samples, as long as you put new, commit into your repository or you can just upload.

F

This is very convenient for local tests. You can upload your contract file here, oops right here and once I have uploaded the contract file. You can see that my cock's has discovered a new, so it's a new API, so I can now find into my API view my catalog view. I can now Friday's a user signed up API. That is an event API discovered from the the syncope. Our contract and I can have the detailed the details of the of the the channel of the operation of mine lock.

F

So I can just have here the different examples, but were in my specification.

F

What we plan to do is to support the the different bindings in the specifications so Fona. We only have focused on Kafka that is configured as the default binding. So as soon as I have uploaded on my definition, my crocks is connected to the the calf calf clusters it is configured with. Here is the end point and it has created a new topic.

F

So the the name of the topic is based on the the name of the API, the name of the of the channel and the version of the API, and this topic can be used to retrieve mock messages that are regularly published by the macaques okay by default. The frequency is 10 seconds that is here, but it can be overridden to two Jet's fit with your use case.

F

So, for example, here when I'm using a default consumer that is provided with the the Kafka broker and if I'm listening to the the topic of the of the G API are just imported, I may start with the receiving new messages. So yes, that's it. So you can see so.

C

It's routed wrap, but we can only see your browser. Okay,.

F

G

F

Okay, so I have to share.

F

Just trying to share my my terminal as well so so here you should see my terminal.

F

So I have used the default Kefka consumer and as soon as I have deployed my specification to my crocks I can start consuming massage like there was a real implementation. Publishing messages on this topic.

F

Okay, so basically, as I said before right now, we support a calf, can message broker, but.

F

Implementation was right now is made to be extensible so that we can support some other type of bindings like an QT or in QP, for example. So we plan to add these. These different bindings in the roadmap.

F

This is pretty cool, so, as you can see, you know that in big corporation in big organization, having a mocking mocking platform that is already up that is already serving but is always serving a P in real time when the specification change, it's very, very crucial to be able to to paralyze the different development teams to be too low, for example, the the mobile application or the the consumer of an API to start working, even if the API implementation is not not yet ready.

F

So we all know the same kind of stuffs with a sank, API.

F

Okay, maybe yeah I'm gonna get back to the pros of sharing.

F

G

By the way, maybe something we didn't we didn't show you guys as well is on the next release. We are also planning to to open what we named my Crocs up, which, which can be a sort of repository of all mugs, ready to use for any type of usages, so that could be for open banking and any application we're starting to build a pretty strong ecosystem.

G

You know with people like Ashley co-op and some others would like to provide ready to use mocking for for it just for four billion users or customers and sound, because that can really speed up. You know the adoption of a solution, because so far those guys well I mean especially their customer. They were forced to install a full solution just for testing purpose, yeah, that's very costly, not free edge I'll and- and we use a lot of resources for that.

G

Thanks to that, we're going to do I mean if we can just enable and lunch whatever they need on demand and doesn't consume that much it's all running on top of Cuban, Tizen and I mean that's, that's very easy. Yeah, as.

F

An example we are starting working with other said before financial services, and we had some samples for all the the open banking, a related specification, so importing them in my cock's is as easy as activating an importer just because I want to get first. Yes up, it has been imported and getting back to my list of API and services. I can see that hey I've got a new mock here, and that is the open banking account. The transaction specifications with.

B

F

The different operation from the specs being there with the different samples coming from the spec and allowing user to to understand the spec and to play with it, we've realized samples, so my cross, really. We want it to be a very easy way inconvenient way to build sand boxes to play with API.

G

Yeah, so unless it's a back to async, API, just very happy to be lets, say challenged by you guys or the QA session to see what do you think about it? If you guys think there's something that may be interested for for yes in API uses and command for the community, and last but not least, if there is any improvement in cement or ideas, you would like to share with us going to be happy to to think about it. Of course, yeah.

C

So I actually have one question: is it's only the messages can only be triggered by frequency like every X seconds. Is there any way to trigger so.

F

Right now for the the first release, I think it will only be by by frequency.

F

One of us feature we have already discussed by other users is the ability to to trigger the publication of mod message with the invocation of mock API endpoint. So basically, what we'd like to do at the end was when you invoke the, for example, create user mock endpoint that is published. That is exposed. My mic rocks you hand, approving user, signed up event published.

F

The Koshka topic, for example, but is also managed by and by my quirks so I, don't think it's will be in the next release, but something we have in target. That's.

C

It that's actually interesting feature: I was thinking they're like master most of people, don't have only event-driven systems, but have a mix like rest or fql with event-driven architecture and and I was thinking that it will be super useful that you post to slash users and then it it injects a message in half hour and QP right. That's that's! Pretty cool I love it. The other thing that I had in mind while watching your demo yesterday is you I, see you're relying too much on the examples to generate in the sample data.

C

Just pretty cool and I love it because it gives you the ability to have like custom custom examples, not just auto-generated and meaningful data right there, that's pretty cool but, for instance, I I, think I, don't know if you familiar with the extensions it's the same as in open API extensions, those starting with X, okay on the spec I think it will be cool if you could create something to create some extensions: an async API yeah that, for instance, in my head, is like X minecraft, there's frequency, and then there you can tweak the frequency on the document itself.

C

You can overload your configuration and on the on the product right and and say, for instance, with maybe then the opportunity to customize how you build the topic. Name. Kafka, like you, don't use this long string right, but every it's not like something else, as a prefix, right and and yeah I mean also for those who don't want to provide manual data on the examples. It will be also cool that you go field by field like, for instance, in the IDE user ID you specify X my Crocs.

C

Random string right and then specify 32, whatever I'm, just imagining, and then on the on the sent. Add you specify that X? My crux now write that that would be super cool because then you can. You could then have the functionality to generate like random information without having having any examples right it's right now. If you rely on examples and people rely in the same document for documentation, they will see your syntax on the documentation. You know my crucifix impacts on their documentation and, and you.

B

Might not like.

C

It right that's why I was thinking that these templating syntax could be moved to extensions, so so documentation remains clean without any microbes reference there right and I'm saying, because today is just microbes, but tomorrow there's another tool that relies on examples not for testing, but for something else, code generation, maybe and and then there's a conflict right, yes, know how to specify the examples, so so yeah I think that's a good point here.

C

Then that would be interesting to see ways to configure with also ways to configure this, the bindings, the the frequency, how to generate the mark-up data.

F

We've got also something specific and I was thinking about extensions for that. It's what we call the dispatcher the dispatching rules, because once we you import your specification artifact into my crocks, we analyze, for example, requests and responses, and we infer matching rules, but you can override this matching rules so that you will be able, for example, let's say to provide right here some kind of scripting here it is a groovy script that will allow you to have matching words that reflects, for example, different business rules.

F

Oh and I was just thinking about using a specification extension for this dispatching words, but it will also very good idea, as you say, to to use it for a specification of random values in examples, for example, yeah, that's actually.

C

Pretty cool example yeah you can inject code to manipulate, for instance, the payload of the message and, depending depending on the payload, you generate one kind of mock-up or another yeah.

F

So this is something we we already have. You can, for example, specify some part of the offer response of an API using some content provided within the request, so that you we are so that you get a dynamic content generated, but yes, definitely I think we should. We should try how to do this using a specification extension. Yes,.

A

One question for my say this, like notation, that you use in the examples like this now function or random. So is it something like specific for you or it's a standard, because I thought that you said it's a kind of standard.

F

No, it's not kind of standard I. The the kind of the template in China that is in my crocks is specific from my crocks, but I just I used this notation with a double mistake, but we can find elsewhere, but this is the name of the functions, for example a random string. Now this is something that is specific to two microbes. Okay.

F

Indeed, I I was looking for something very lightweight and very performant and did not find something great in what was existing so I.

C

Wonder if the fall except postman have something like this, that they can open source, maybe will be interesting to talk to them. Yeah, yes and I. Think it's the same syntax right! Yes, it's very close, yeah, very close yeah, so wait to talk to them yeah. If there could be a potential collaboration did but.

F

What I've seen in the postman is fond of the simplemente tin, know, GS and basically.

F

My cock's is mainly Java.

C

In Ghana, I'd meant more, like you know, following the same syntax standard way of doing it, and then this print could be maintained by all of us right, not just microscopes postman, and then from this specification you create implementations. Orientations right sounds like this is a really longer, but I'm, seeing like like Lucas was mentioning to me. This is familiar. It's a familiar way of describing things and I'm sure Mike Rutzen postman are not the only ones using this syntax I'm pretty sure.

C

So it will be great to have some efforts on syntax as well outside.

C

G

C

G

Has been introduced in the latest release and we didn't see any start out for that, so we try to let's say to select something which seemed to be pretty well known and common on on multiple project and to tell you the truth: I mean we. We have had a lot of conversation within the community for this new request, because at the beginning we didn't want to generate any random stuff.

G

We like to force users to to write their own and real example, but there's a lot of people within the community who said now: I I, don't care about that and just a developer I would like to do it quick and dirty I. Don't care and it gonna. It's gonna really help me. So that's why we honestly we have created that it's it's due to the community and based on a lot of requests we receive on that. So you just do the thawed-out point. It is important and you're right, that's something we need to defer.

C

Yeah also also doing it manually can be kind of painful for certain data types like like.

C

Every time I have to ride one of those manually, for example, I, go to Google and search or I, go to the browser console and generate one cuz. It's too much and it's stupid because it's actually pretty simple syntax, but it's so it's super long and it's very error-prone, so yeah I understand for.

A

C

I left a link on the chat today, our extensions catalog at some point. Hopefully this year, we're gonna create an extensions catalog for async API. So all all the people, creating extensions with excess, could be listing their extensions there on the catalogue. Think about it.

C

Like a public listing all the old extension said, you could use witnessing KPI for different purposes and, and it also allows you to define validation rules of your extension like it needs to be structured this way, so it's like kind of a mini suspect only for the extension using decent scheme and and then next next step will between two great this extension catalog with with our parsers. So you can opt in you.

C

Can optionally check you can optionally tell the the parser to go grab the definition of these extensions and validate if the data defined inside extensions, it's correctly formatted and it tests correctly. It's been correctly used so right now that the parser ignores everything that starts with x desc.

C

So it's not validated, but we see some certain value on having some extensions that people can use like in this case, for instance, it would be great to have a set of extensions for microbes or maybe one, but so that you can define all this specific data for your maken and and yeah, and you can have our parser to validate that what people are writing on there. A syncopate documents is actually microts valid information and correctly structure, so yeah and the release cycle will be different.

C

Test is different than the spec, so you you will control the release cycle of your of your extension. There's a there is a bunch of ideas there on how to how to make it possible. It's just like a draft project we're just starting, but it's pretty much I will say in the past to become a reality. So so again, this is a kind of way to standardize, custom stuff, so so yeah. Just to make an example, one of the common extensions that people who find there is you see that nascent KPI an open API.

C

You have the info object with title description and and then you have contact with email and all the stuff right, so contact could also be extended with Twitter, for instance,.

C

Or a link on reddit or something for support, so this this extensions could be useful for many people to reuse instead of everyone creating their own and eventually, if it makes sense, they could be including two into the main spec right so yeah. This ability to know about.

A

Few minutes left and I have two curious questions like not related to us, an API but I'm from my say, I worked in ACP before and there was pretty really specific standard there. But it's it's an open specification. Do you have any requests and are you going to support? Oh data or data.

F

Any requests about supporting your data for now, okay, just.

A

Curious because we had to like when I was working in the previous project, we had to build our own mocking tool for OData, okay,.

C

A

Was nothing, and so in case you're gonna, think about it and we'll have requests, you can contact me I know! Yes that developed it and they can give you hints okay, cool and technically how? How does it like? Because it's running it on on kubernetes yeah? So you basically every every API? What do you do? You basically generate a an application that you'd, like you, spun a pot or a door.

F

And that's it or no indeed,.

F

And that is something that allows also the scalability. There is no specific endpoint that is generated for for an API mock. Everything is dynamic. So, basically, when we deploy on Cuban entities, there is a bunch of the pods components for the different features of the of my crocks, and we just have the autoscaler to manage the number of replicas of the main, my crocks instance to Lotus k2 tens hundred millions of of API.

A

By dynamic you mean okay, so you're, it's one application, but you just fit it with different spec and then yeah. So.

F

And basically,.

A

F

Are using the operator deployment model to be able to easily deploy all the components within my crocks and also touched to change the topology of deployment, depending on the load on different services on different endpoints? So, basically, what we? What we on the pass is when there will be a lot of, for example, a sink, a pi/2 to simulate the the operator will be able to registry beaut the load between different pods, but will under different publication of different API.

G

It has really been made to be highly scalable and to to to support a huge amount of requests. Yeah I.

F

Was talking earlier about my experience in a larger financial services firm, and we made experiences about deploying mock using a software engine, for example, and at the end of the day, we realized that the infrastructure we have to set up for only the mocks was as huge as the infrastructure we are introducing for the ESB. So really it was not something that was scalable and my course was build from scratch. From this experience to have something that is, that will scale to billions. Hundreds of hits without.

A

F

A

So thanks a lot I mean I, think it's already ang at the meeting, so guys awesome demo and it was really cool to have you here and see Maddie told in action, although I've personally watched it before midnight yesterday, instead of going to bed, I was watching the demo, but still I was the some to to hear the technical details, especially for me. So thanks a lot for for joining front. We want to add something at the end, not.

C

Really much I mean thanks for for yeah. Let us know if we can help anyway in the future. In the next days before you launch this, you want to test something or to review something and we're happy to collaborate. There.

F

Thanks a lot for having us for giving us the opportunity to play the demo.

G

Thanks a lot and we'll stay in touch in any case, because that's really something which is expected by the some of our people within there or you know the micros community so happy to to share an update regarding the latest news on a topic.

A

Awesome, that's good.

C

Cool anything else once and if anyone else wants to add something.

A

Three-Two-One thanks a lot and see you in two weeks, but this time in two weeks, of course, in a pack friend any time, oh yeah, hey guys.

F