AsyncAPI Community, 22 Jun 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: AsyncAPI SIG meeting 51 (June 22, 2021)

Description

This is the recording for the AsyncAPI Special Interest Group (SIG) meeting #51.

https://github.com/asyncapi/community/issues/31

Chat:
00:21:14 Sergio Moya: Never disturbs :)
00:21:24 Marc DiPasquale: All good!
00:27:41 Fran Méndez: brb
00:27:43 Lukasz Gornicki: https://github.com/asyncapi/spec/issues/536#issuecomment-865898634
00:41:14 Kin Lane (Postman): You just did
00:41:30 Marc DiPasquale: So Kin is going to break everything?
00:41:40 Kin Lane (Postman): Yes! That is my job!
00:41:41 Mike Ralphson: No change there then.
00:41:44 Marc DiPasquale: lol
00:42:04 Lukasz Gornicki: https://www.asyncapi.com/blog/may-2021-at-asyncapi

A

Thank you, bot, um hi folks. This is.

A

I'm not recording podcasts. Okay, hey uh welcome to asking api meeting the the usual one that we have every two weeks. This meeting is being recorded, so just be aware that you're gonna be published to youtube.

A

um We have three topics on the agenda and looking on the list of people that joined the meeting, nobody knew on the meeting, but before we jump into those three topics, anyone has something like super important to discuss now, so we can leave the meeting and don't you don't have to stay for the entire meeting.

A

A

Fran, you want to say something about your king x-king, no.

B

Okay, it will not happen again. I promise okay.

A

um So there are the the first topic. Is actually you front, so I'm gonna make a co-host. If you wanna share, do you wanna? Would you just show something like sure.

B

Yeah I'm gonna show my screen join my screen.

A

So I'm just yeah front will continue. The first topic on the agenda is requested by mark hello, mark and uh hello, mike, hey friend,.

C

Hey lucas yeah.

A

Thank you. It's uh it's a question um from mark like to, if it's possible, to get some updates like on on the roadmap, where we are um and yeah front on to you.

B

Thank you so well thanks man for bringing in the topic because yeah this this road map is it's great to have a road map, but not great. If we don't update it right here or we don't update the status. This is something that I promise I'll find some time soon tm trademark uh to to improve it.

B

So people can see real time how how we're progressing here, um but uh in the meantime, I'm just gonna take manually, I'm just gonna, give you an update and uh so you'll know where we are right now, okay, so so the now call you were, you were asking for that now column.

B

So, what's going on on the now column, okay,.

C

uh Yeah so yeah, so just I guess a little more context right like I can't you know, I I'm obviously not plugged in day in and day out to what's going on in async api. So I saw this roadmap. You know a few weeks ago, and then I was digging into it the other day- and I was like hey like a lot of these things- seem really cool. I don't feel like I know.

C

What's going on with a lot of them, I know maybe some bits and pieces about some of them and I'm sure there are a lot of other async api community members in the same boat as me, like we usually keep up with like one or two or three things that we're most interested in um so it'll be good to hear.

C

Overall, um you know where we're at with the overall roadmap, and so I specific specifically called out now, because that's obviously like what you what it's currently working on and what we can expect next, but depending on how much time you want to spend on that, like obviously I'd love to hear about like what next priorities and stuff are as well yeah. So.

B

Okay, so so in terms of priorities, uh I think it is what it is there like. What you see there? It's, um we don't have like a in order of tasks. We actually don't have tasks. So what you see there are not tasks. This is a a it's an outcome-based roadmap. It's the the the name of this, the name of this kind of roadmap. So it's not a like a sequence of tasks.

B

Another way to see it is that if we expand this things like solutions right, the ones in blue at some point, whenever this is really finished, you will see a lot of tasks listed here like the ones you see, you're, seeing here on the on on gray, that gray or even black people say, um and that's where you can see the task okay, but this roadmap is um to look at the future a little bit above tasks themselves right. So it's about outcome.

B

What is an outcome, so let me just check if it's here yeah so now it says an outcome is the result of implementing a solution and represents a specific milestone towards our goals and therefore our vision, if you're into okrs an outcome, is a clear result, it's the same thing. Okay, so this is like you know, kr's. This is a killer right, so um these are like to summarize. These are like milestones that we want to achieve to arrive to our goals right, the three goals that we have listed here.

B

Why? Because, if we are, if we get into these goals, if we succeed in um accomplishing these goals, then.

B

We'll satisfy division right, that's the that's the whole point of this right, so this goes like, like a inverted funnel. If you want so it's like division starts, then it starts with three goals, and then each goal uh might have different uh different outcomes. An outcome usually is based on on user needs on user problems. That we want to fix in this case, usually is developer problems that we want to fix.

B

So, for instance, that's why you see that's why it might read weird at first, if you're thinking about in terms of tasks, it's like yeah. This is too generic right uh for a task. It's too generic, of course, and it's not really defining. What do we have to do? Instead, what it's defining is we're defining the result of a bunch of tasks that we'll have to do in the future right. So where do we want to arrive right?

B

So, in this case, is user is able to focus only on business logic? um For this to happen, if we want to make this a reality like the user, the developer only have has to focus if the developer wants, of course, only in business logic.

B

There are many things that need to happen right, but at least two of them is that we need to. We need to improve generators, it generates data models and- and we need to make generator support- react because yeah.

C

B

How we improve generator to make uh to make better templates, but there are, there are other things not just that. The confusing part- and I agree with this- is that this is not a complete list of things uh that we'll have to do right. So you'll see things popping up here uh over time in here.

B

Why and while they are not defined here, we don't want to. We don't want to um like very upfront. We don't want to be doing this. This work of we're going to do this this this and this because what happens usually, is that whenever we do it- and I can- I can confirm that everybody doing this in the whole world.

B

It's all. So, if you plan for a roadmap beyond, let's say four six months, it becomes most of the times because you don't know how things are going to change. You don't know how things are going, um things that you do in the way are going to change the priorities at some point are going to change, not just the priorities but they're going to change the way you now look at the future you're going to solve problems differently.

B

So we don't know what's going to pop up here, but as soon as there's something popping up that is related to improving this outcome, it's going to be listed here right, so the only thing that we can, that we can ensure or assure you here- is that we're working on this we're working on the on this outcome and because it's in now it means that it's not finished right, and it also means.

B

Let me just shake him because we have people joining okay, um I think, are you co-host, because I I get the request of people joining? I don't know if you see it.

A

Yeah, I am admitting, I also see.

B

A

Yeah you can ignore it. I will admit.

B

Okay, so um so, as I was saying like this is actually where we're focusing all our efforts, we're focusing on our efforts and making the user being able to focus only in business logic right and what are we going to do to make this possible.

B

We decide on the way right like each time we decide. We agree on issues depending on the discussions and the issues we keep discussing. What's the best way to get there right, so that's it that's what it means.

B

Similarly, it's impossible to unsync documentation from code, it's also very related to um to how the like it's very related to the to the other. One, like people can focus only in business uh on on business logic, but in this case it's more than um the developer aside from not forget not having to focus on on on non-non-business logic. What happens is that we kind of we want to automate it. So we we want to solve another problem which is uh okay.

B

Imagine that I can only focus on the business logic cool, but now still have the problem that everybody has that we have the condition on one side. We have our code here and they both tell different stories.

B

So you deploy your code and it's doing and it's not doing what the documentation says and uh if you have documentation right so so we want to solve that problem as well as much as possible, and for that one of the solutions is to create a framework which I'm working on right now and um and another solution is to create plugins for the most common frameworks like um what we call this java spring cloud streams, for instance right so create amazing, kpi plugin for screen cloud stream.

B

If possible, we still have to investigate right so in a way that um the code will always honor. What's in the json kpi file, and if it's not in this npi file, it will not work right so and at the same time it will generate the documentation from this api file. So that way you always make sure that, what's in the code and what's in the documentation, it's bold both are in sync right. So these are different ways.

B

uh Stop me. If you have questions, I don't know, I mean any of you not just you mark, so so yeah um I mean each of these items here. You have this, you have descriptions right so and we can discuss there so happy to follow discussions. There.

B

um Well, so that's another another outcome that people are worried about. You know which is like. um I don't know if you're hearing so much the baby, it's disturbing. Is it uh yeah? Okay, sorry, look it's magic! If you depress.

B

B

Was saying that um this is another common problem people facing this problem that they want to make sure that whatever ends up in the broker, it's according to there snk definition right. We can do this in many ways.

B

For instance, one of them is with the framework a framework can validate the message before it gets into the broker, but what, if you're, not using the framework or what? If you, what? If you already have some code base, which is that's probably the case for big companies right um in this case?

B

One option is that we create an event gateway right, so this validation will happen in a infrastructure piece that will be limited between the client and the broker, and the only thing you will have to do is to install the install the event gateway in your system and change the dns right, so your broker is actually the gateway and then from the gateway. You forward messages to the broker right. So in that way, clients don't need to change their code.

B

It's only a dns change right, so yeah, that's uh one solution. um Look at this! The vision and the roadmap are clear. So that's another outcome that uh we want to. We want to improve, and um this is done. I mean this landing page is precisely this thing.

B

But it's not done yet. The whole outcome is not done yet and you're the proof because you're asking about it. So we need to clarify this, so we need to make it clear right. So what's next, uh what do we have to do to clear to make it clearer happy to to take uh suggestions for sure something that we're gonna do at some point? Is we're gonna start listing all the we were doing an experiment before that's why you see some some things here uh and not in all of them, but now we're gonna.

B

Now we're gonna um integrate this with all our github, repos and and issues, and you should start seeing all the github issues at least yeah.

C

I think that I think that'll be really useful yeah. I think that would be really useful, like kind of as you as you go through this right, like you you're talking about a lot of like what we can do, what we could do, but it's hard to know what is actually being worked on right now right unless I go click on these things like, for example, when I click on the create an event gateway looks like somebody is working on that and you have a lead listed there.

C

But then I click on some of the other ones, and it's kind of you know we could create a plugin for x framework y framework z framework. But like is somebody going to start with x framework- or you know like like that like like who is working on or.

B

Is nobody working on it? It's just.

C

B

Yes, actually yeah, that's probably confusing, because what we move from the thing that you can move from columns is actually the outcomes and because this is this is part of this outcome. It's moved to that column, but yeah uh nobody's working on this. Yet that's okay, yeah yeah.

C

So I think it would be useful like if you expand the make generator support, react right. It says the task or the feature below that says done. So if, once you have the features there, if some of them say done some of them say in progress, some of them say not started. For example, I think that will make it I'll.

B

See some some progress indicator here like okay, so this is done or this is not even started or yeah. That would be great yeah, okay, okay, thanks really helpful, um so yeah and and then the that's another part, which is some many people. Many companies already have existing documentation and developer portals and that some of them they created on their own and so, instead of just generating the whole page for your documentation.

B

What we want to to offer here is just chunks of this page right, like some people only want the message view or the list of channels or uh to integrate with their documentation right. um So that's uh for that we're going to create we're we're actually working on it. I think magic is working right now, so so yeah, it's a ui library.

B

Instead of a single react component, we split the react component in many react components, and then it's not only for react, but also for other frameworks and uh and so and so on I mean like this is what's going on in now.

B

uh In later you have things like you know, like specification and parser are ready to accommodate future features and other stuff I mean.

B

If something is not clear, I would suggest go to each of this and leave comments and we can have a discussion there if you want to add something to the roadmap, because you think it's missing something like this is a road map that we all do together, so so yeah feel free to suggest.

B

And that's it for my side and questions.

C

Thank you, fran appreciate you going taking the time to go through that. No no problem. Thank you.

A

No other questions to from okay, so the next topic on the agenda is released, one of the spec, so, first of all, I'm gonna share with you a link to a repository where we track the progress.

A

uh To the um to the github issue, so we are now uh I mean we released. I don't remember, we are pretty pretty productive with this pep releases. If it comes to release candidates, we released something today, which is.

A

um Release candidate number six, no as uh number four for the two one release, um because we merged today a change from from dale to add ibm mq to the list of the bindings.

A

So looking on all the um all dprs and their status against the contribution guide, uh there were two additional candidates for one: one is the endpoint mq, but it's more complicated than ibm mq because ibm mq yeah. It was easy to handle because the the binding itself was added by by richard from ibm. uh I think a few months ago already, so it was just about adding it to the list in the spec with any point in queue, uh it's more complicated because uh the binding itself is not yet merged.

A

um We left some comments with run for generals um and, let's see if, uh if then these can. These points can be solved anytime soon, before two one happens um and last but not least, it's something that is now super active. It's a it's a proposal from lauren um about extending the the examples object in the message.

A

I will extending it with two additional properties: optional properties. um So the proposal to the spec is already approved.

A

uh We're just missing the um approval on the change in this json schema for this pack and also the parser.js. That's an ongoing discussion on the pull request, but these things are trivial so, like I'm not going to make any promises, but it really looks like this. This change is going to go into 2.1 release.

A

If it comes to the automation, I use it's just doing some fixes on the website side, not blocking the release anyway, but looks like for this to one release. We're gonna have like full automation that, like, after releasing the spec, um the document will be automatically published to the website, which is super cool, um so yeah it's scheduled for june, and we hope to not do any release like last day of june.

A

So um if examples um are solved soon and we clarify what's happening with any point in queue um pr, then hopefully 2-1 will go into the wild this week.

A

Any questions from your site regarding 2.1.

A

Oh ok, cool and last but not least, uh what was it?

A

Yeah uh just wanted to point out um like the next topic on my personal list of things to do for for the community, and this and asking api initiative in general is uh to make another event about asking api. Like the last time we had last year as an api conference, uh we need to do something this year.

A

um Some time ago I talked with ron that I'd love to have some hackathon, um like the one I participated in with github actions that you have like month. Long hackathon and people like can hack for for an entire month and then show what they did to a committee, and then we just choose the best projects um but, of course, an even like classical event as a conference, that's also an option.

A

If there's any interest, I'm happy to work on it. The question is to you like what anybody has an opinion here like about some as an api event.

D

A

Many people say it it's a lot of time, but yeah I mean uh it's, probably that's something that we're gonna have to do. um Anyone already knows that you would love to work with me on the on organizing the the events or you just want to participate. As I say as listeners on it's pretty cool to work with me.

A

Not many people worked with me in the past and now some are still alive.

A

You can share it in private later. If you're shy, that's fine, but did you did you ever attend um or took part in in a hackathon, because I in the past, when I was doing hackathons, uh I was participating in hackathons. These were usually crazy weekend hackathons like several hours, um not something that, after that, after the hackathon I enjoyed.

A

uh If it comes to my my body and uh exhaust exhaust exhaustion um and last year I saw this hackathon for the first time, like you have mont long hackathon did you do you have any experience with it like is, it is? Am I just biased by my experience or.

D

That that was my suggestion for the when we did postman galaxy last february, which was three days event. uh We did the entire month of january as a hackathon and the first day of postman galaxy. We announced the winners of the the hackathon, and I thought it was a great lead up to um you know it kept us talking about the hackathon as well as the conference, all the throughout the month of january.

D

So it is. I agree. It is much more work um which I'm I'm here to help you with, of course, but it they they tend to compliment each other.

A

Okay, nice, I didn't think that you can actually join them. It makes sense like you, you make a summary on the on the conference.

C

Yeah- and I will say like so- I've had a good hackathon experience for like the in person hackathons, where it's over a weekend. I don't think like over the weekend. Thing will work virtually personally, um so I agree like if you want to do like a virtual global hackathon. You probably need to do kind of like what ken and them did at postman like a full month type thing, so people can do it kind of on their own terms and maybe have a few online office hours and stuff like that. So definitely.

B

A lot of planning.

C

To go into it, but I think you'll get we'll get some value out of it. I'll also say I do think I did like async api conf last year, um and so I do think it will be worthwhile doing a similar conference to that as well. So just my two cents, there.

D

Yeah- and I would I would second that that I think the the conference would be my priority over the hackathon, because I think right now, there's a there's, a lot of need for large organizations to understand what async is and have conversations about it. So um I think async confirm be a great place to make that happen.

A

Okay cool, so uh there is a dedicated channel in our slack called events, so I would like encourage you to join, even if you don't want to like participate in organizing uh but like validating the ideas, the discussions we have there.

A

um That would be awesome if you could join and and voice like, especially that like last last year, not sure if you remember when we were organizing us in the api uh conference, uh all the proposals we were reviewing with um also like independent members of the community.

A

So we definitely will also have to have some volunteers from from the entire community from people that would like to also help us out to choose topics and.

A

And help at least this way.

A

Any more comments regarding this.

A

Anything else you'd like to mention, because we covered all the three topics from the agenda.

C

um I just had a follow-on question. I guess to the previous topic, uh where you said 2.1 is set to be released in june. um I was trying to look. Is there a defined release? Cadence now will 2.2 be in like july, or is it once a quarter or once a month or it's.

A

It's not clear if it's 2.2 or 3-0, but it's the cadence, is defined it's uh it's always gonna, be in june in september, in january and june and september and january. No, no there's something missing.

B

No, it's uh april.

A

April yeah yeah uh june, uh it's uh january april june september. It's always the same. We wanted to avoid the uh holidays and christmas. So that's why.

C

Sometimes it's every.

A

Three months, every four months, but.

C

Okay, cool. Thank you.

B

Yeah- and we don't know, what's going to be released, it all depends on the content of the release. If it's.

C

B

A major or minor yeah yeah, I know.

C

um I know we've we're kind of looking at some of the binding definitions that are there and um I know I think ian was looking at updating the amqp ones and we were looking potentially at putting some work into better define some of the mptt ones um and maybe an smf uh which is the solace protocol as well. So um just a heads up there, but I don't think it'll be in the june release. At this point.

B

For sure yeah we can safely say that this release is closed, but.

A

Now, at least, we.

B

A

This guidance right so like september, is also pretty soon right.

C

Yeah yeah exactly so, I can bring that back to the team and be like hey. We definitely want to get our. You know.

B

C

In a row to know, what's going on before september, rolls around.

A

And we're also and we're we're, of course, talking about major and minor, um like patch releases, they are they're gonna, be released immediately, like bug, fixes or something.

B

Yeah and that shouldn't be a problem because since um if you read the spec in the beginning, it says that the patch version is for non-normative changes right. So that's to let's say that if you're supporting version 2.1, let's say 2.10- and we release two point- one: twenty: okay, uh it's it still should work with your tooling with your tools right, because all the changes on the patch releases so will be non-normative, mainly editorial fixes and typos. And all of this so yeah nothing.

B

Nothing else, and I also wanted to clarify that we don't know, what's going to be on the next release, and but we also we're not stupid right and we don't want to be uh here and uh we're not going to be releasing major versions of three months right.

B

So if we know if we foresee that there will be a bunch of breaking change or breaking changes, uh and I'm not promoting king here um king's show breaking changes. So if there are you just.

A

Did by the way,.

B

But I said I'm not sorry, it's not it's not so.

B

So if we have a bunch of breaking changes that are in the foreseeable future, we of course will hold it we'll hold them right and group them into a single major release right. So you only have to break your code once and not every three months.

B

B

Making changes.

A

Any other questions.

A

So just advertisements from my site today we released a blog post that I wrote about things that happened at asking api in may, so you might like to have a look especially like mark you you're thinking about fixes to the bindings like now. Bindings are not only described as with markdown, but you can also provide like we expect jason's keynotes, for example, so yeah.

A

Nothing else: oh three, two thanks a lot for joining another meeting in two weeks, uh we're starting holidays, so don't expect much, at least for me during this time, um but anyway have a good one: cheers.

B

Thank you folks see you around.