Magento Community Engineering Hangouts, 2 Feb 2023

Previous Meeting

Next Meeting

⏯

youtube image

►

From YouTube: Community Engineering Hangout Feb 2 2023 Meeting Recording

Description

Community Engineering Hangout
Feb 2 2023
Topics Covered:
1. Catalog Service
2. Community prioritization process

A

First, one is a catalog service for Adobe Commerce and the second one is community productive.

A

The second one is community prioritization process, um so we will start with catalog services for Adobe Commerce. First before we begin I just wanted to, let you all know I'm based in Austin Texas and we are experiencing severe winter storm. So you know, power lines are down, uh trees are down, uh it's really really cold outside so in case I lose power or connection. But if the meeting is still on, please continue I just wanted to. Let you guys know so with that I will hand it over to Sandra.

B

Thank you, Carol and don't worry, hopefully, you'll be fine, but we know we are aware that there are some issues with electricity and things over in the U.S. So thanks for letting us know and I think from the list of people that they see on Nepal, I.

B

Think um most of you already know me, but there are a few of you that probably we haven't met before so: I'm Sandra, Gonzalez, I'm senior product manager, working with Adobe, already for some time and I'm now, working with the Catholic Service Board of e-commerce and that's the main topic that I'm going to share with you today and let me just share my screen yeah. So you should be able to see my screen now.

B

Hopefully, okay, so thank you. So, yes, I'm gonna be talking about the Capital Service as I just said, and I just uh oh I'm, just starting with the last slide. So maybe it's ready to start with the first one.

B

uh So I'm also going to be presenting together with lotion, who is our senior computer science and it's going to be taking care of the demo of the capital of service and a little bit more technical details, and we also have press done that is the senior Union manager and and will be supporting the conversation and in which case this feel free to ask questions over the chat.

B

If you, if you have any doubts or any ideas, he can support us during the this conversation, and so first uh um I think some of you might already know of more about the Catholic service. But I just want to make sure that we are on the same page and give you an introduction of what the catalog service is.

B

A service is one of the SAS Services thing: that's what we have with live, search and product recommendations. So it's independent from our core from the Adobe Commerce core, basically augment the apis that we have with the the catalog graphql apis or the core apis that we have at the moment, and the idea is that returns a product data much faster than what we have with the core apis and when I say much faster.

B

We've seen in our testing and I just want to highlight that that's what we've seen in our testing, but the performance Improvement depends in a lot of factors: the the environment, the data, the catalog, the settings that you have on each of the projects, but on our test, if we have seen up to 10 times faster Capital data, retrieval so in in responses under 200 milliseconds, which is a very, very good comparing to what you might have seen in in existing projects.

B

uh So that's the the one of the big improvements and benefits that we are offering with the capital of service, and we I also want to highlight that we are. We support, B2B use cases with share catalogs. So that's something that is a out of the box with the capital service.

B

From the beginning, since we launched the service later last year in October, then because it's a SAS service and it's a double managed, so it is scalable, it is cloud native and uh because it is independent from the core difficult is that it also lowers the traffic on on Commerce service, which should help to improve the performance on Commerce Service as well.

B

The the catalog service has the possibility to work by itself as a standalone, or it can also work together with the other services. Basically right now for recommendations and live search and I'm gonna get into a little bit more details into that. But it's it's the idea of all our SAS services that they complement each other and they improve the experience.

B

It goes for the developers and for the end users for all our Commerce customers and when I say oh I've got my customers I'm, also implying that the catalog service is only available for Commerce merchants and it's it doesn't. It doesn't require additional, it's not an additional price, so it is available at no cost.

B

Okay. So- and this is an a high level overview of the, how is the capital service, interacting together with I'd, offer Commerce monolith and with heads at the storefronts and other staff services?

B

So what you can see here on the on the left of the diagram? You see core graphql apis, that's the graphql apis that we have already available today and and that, inter, if they are all integrated with Adobe Commerce and they can be integrated with different uh heads. Pwa is one of them, but it could be any other front-end that the merchant is using and I.

B

As you can see here, we we are showing the storefront latency, but this is the time that it takes for the core to respond or to to to to to to retrieve the product data to the front end- and here is where we.

A

Yeah sorry to interrupt, are you sharing a screen.

B

I thought it was, but it looks like is it? Can you see it? Yes, we can see it yeah.

C

We can see that thank you.

B

C

Sorry about that.

B

I was not sharing anything very interesting before so. This is where the fun starts. uh Okay. So what I was uh talking about? Is this storefront latency as we call it? And basically this is the time to get the product data uh on this front end on the storefront from in this case, when it's through the core graphql apis. It comes directly from Commerce from from the monolith from the Adobe Commerce model, and this is the latest. This is the time that the catalog is service.

B

It's improving significantly and the way it's done is that we have our SAS services. That is this uh bubble that you see here, and our fast services are indexing data. In this case, we are indexing, product data and storing product data within the capital service, and then this data is already calculated, so we are not doing any calculations like price calculations and anything that it. It is something that is consuming a lot of time when you are asking for this data directly to the monolith.

B

So that's what it helps to get the data much faster, because it's it's also an improvement on the technology that we're using. But it's also that it's getting the data simplified much easily and you will see also on the. What is that we're using how we have improved and simplify the the queries that we we have on the cardboard service to simplify the product data retrieval as well, also something that uh it it's great to have, together with the capital services that were stored from Gateway.

B

What the storefront Gateway does is that it federates all our SAS services, so we're still working on program recommendations, but it is already available for live search and what happened with live search is that it's amazing. It's an intelligent uh enjoying that returns. Amazing results based on the criteria that nurses have defined uh that, but it has some limitations as well and the limitations that it has some of them. We are complementing them with the catalog service through the cap, the storefront Gateway.

B

So what we have done here is that, when you're asking results through product search query, you are getting the list of products from live search, but with the product data from the cabinet service, which is retrieving more data that would live search is has at the moment and when I'm saying more more data, I'm referring to custom product attributes options and we're also planning to include color swatches.

B

As you can see here, you can use the the store from Gateway API directly to connect with any front-end and there's another option. So the next option that I'm going to show you is the possibility to use the capital service and well also, potentially live search and recommendations. Of course, through the API match, the API managers you know is it allows you the capability to integrate multiple systems, and we I have just highlighted here just what it relates to Adobe Commerce, but you can also use it to integrate with any other third-party system.

B

In this case, what uh we are providing I I mentioned at the beginning, that the Capital Service augments, the the core graphql apis and the features that we have right now. So our intent is not to replace all the features that we have today within the choreography or API. The main reason is because we're focused on the features that provide better and higher value to our customers.

B

So, but we don't want to limit the possibilities of using the catalog service because of some features might be missing on the implementation, as we are not planning to have 100 feature priority.

B

So the idea is that, through the API mesh, it is easy and possible to connect and to get the data both from the catalog surface, with a high in performance. Improvement and other data are not available apis within the same query.

B

So the idea is that we are complementing the performance improvements with a specific features that are not available on our on our SAS service and one example is if the the product entity ID, that we don't have it within the capital of service, and we are at the moment not planning to introduce it into our query because of the risk and the complexity of adding that. But it is possible to get it through the API mesh connecting with the coreography Olivia.

B

Another option is, for example, another case is that at the moment we don't have support on Capital service for tier pricing. It is something that we have on the Roma for this year, but it's still not available. So this is something that it could also be implemented through mesh using for graphql apis.

B

And with this I want to share a little bit more of what we're planning to do also uh on our Roma. What's our roadmap looking like this year and so the first we're working on the category API and the idea is that the category API will allow our Merchants to render the top menu navigation, but it will also be used by imerge to improve the rules that they have so the capabilities of the intelligent merchandising that we will be providing.

B

Then we are also planning to have a product detail page microphone and which might sound special or different challenging. So the idea is of the microphone tent um is to align with unified storefront strategy that we are moving forward. We are on a very, very early stages at the moment. That's why we are considering what the best option.

B

Sorry, I'm still am I still sharing my screen, because it's telling me a message here on on teams that I stopped sharing yeah.

C

Yeah, we can still see yeah, okay,.

A

It's working yeah.

B

Okay, thanks, uh okay, so, uh okay, let me just go back to the protein page marker from them, so our plan is to uh go take with unified storefront, but we are at the moment, analyzing. What's the best approach for us, because we have been working, we start working on this micro content. It's basically a piece of content that will help our Merchants to render the product detail Pages using Capital Service apis very quickly.

B

It would simplify ways of developer experience, so allow allowing you to customize it and to have the design the merchant wants to have. So we are now uh planning and deciding which, if we want to go and deliver something quickly, that our customers can start using uh very soon or that if we we want to wait for the whole unified storefront strategy and deliver something together with all the microphones that other teams are also planning to work on.

B

uh Having said that, the the idea of the microphone 10 is that it will be agnostic, but we are planning to adapt the microphone 10 to each of the main front ends that we have and right now we're focusing on Luma and am so they with this. Our intent is that Luma and AF customers will be able to adopt micro front end and start using the catalog service very easily, as I said before, we are working and planning and designing how to approach this, and that's why I'm calling for partners emerges.

B

So if you feel this is something that resonates with you and you're interested in knowing more potentially to collaborate with us feel free to drop an email to to my inbox, I'll be more than happy to to talk with you about this, and then we have we're also planning to improve.

B

This is not a within the catalog service, but it's also that I know it concerns a lot of our customers, that is the size price indexing uh so that it impacts the time to import data from Adobe Commerce and to make it available to the storefront.

B

So, with this initiative, our idea, our intent, is to improve the performance of this data injection. So we will reduce the time that takes from the moment that you update a product or that you import a product or multiple products within Adobe Commerce and those products get available into the storefront.

B

Then we have the possibility to enable prices for stats on Luma projects, and the idea is that for this initiative, what we're planning to do is to allow Luma customers to decide whether they you want to continue using prices from the monolith or you rather use prices from SAS, which will help you improve the performance of yourself and and then it is by the way not listed on any specific order.

B

But it's it's just showing you what we're planning to work on and we are still defining uh what's going to come next in the next few months, and then we have the catalog sync dashboard at the moment. This only works for product recommendations and it basically it shows how data the status of the data synchronize between Adobe Commerce monolith and pro recommendations. So we have.

B

We want to have this visual representation of the status of the data, also taking into account Capital Service and then I'm, just gonna jump this, because I think Lucian you're gonna go into the details to tell the queries that we have uh just want to highlight that from Seven prototypes that we have within Adobe Commerce.

B

We have simplified that to two different project views that that are, those are simple products and complex products that will map to different uh product types related to Adobe, Commerce, I'm, Gonna Leave it now to Lucian and I just want to show this before we jump to the demo. uh In case we don't have time once we finish with the with the demo.

B

But if you want to know more about the capital service, you can join the communities, Cloud Channel store from services, or you can write an email both either to to myself or to Russians and now all to you, Lucian.

C

Thank you, Sandra, okay, so I'm gonna give like a brief overview of the of the queries that you can do with the catalog service and also like uh how the data looks like uh for this. We have been like a super simple react: application, it's running on my localhost right now and that is consuming the uh the catalog service available at catalog, survey.adobe.io and uh I'm gonna work through uh each each query as they happen.

C

uh Okay, so, first up like Sandra, said, the the catalog service exposes like three types of query. uh You can do like a product find based on the skew. You can do a refine which I'm going to talk a bit more, uh as the demo goes on and also it offers since catalog service is a Federated Gateway.

C

It's also integrated with live search, so you can do live, search, queries from catalog service, but with enriched data.

C

um Okay. So let's, uh let's move on so first of all, this, like this is a sample PDP, that a developer May build using the catalog service. So as the page loads, uh the client can do a product query to fetch the the details about the about the product. So uh you you can get information like uh the student name, description images, then you can also filter images by rows. You can get the attributes uh and you can also do a filtering based on attributorial in the future.

C

We also introduce a pagination for attributes in case there are a lot of Articles being returned and, like Sandra said, we have simplified the the model. uh We only have like two types of product now: the simple product and complex product uh like a super quick definition, a simple product, it is any type of product that that does not have an option. So, if anything has an option, it's complex product, if it does not it it's a simple one.

C

uh Okay and uh also uh complex products can have uh price ranges, as opposed to like a fixed price and uh for example, if you have like a grouped or bundle product, then the option is an actual product itself, and this is how you can actually get the data for that product.

C

Okay, so as the page load, the the client will do like a a product query and it will get information uh about the product images options like colors and sizes, uh description and the custom attributes or dynamic attributes uh very important. We also support the B2B and, for example, you can um disable the add to car button.

C

You can hide the prices or you can hide the product all together uh that next up, let's say a shopper, decides to uh filter by or to select an option and, let's say, for example, it selects the the green color.

C

We also support the images based on on the selection. So when a shopper selects an option, then a refined query is uh is made into the the catalog service.

C

It has like two parameters: one is the skill for the product that you uh you're doing very fine, and then you have an actual option of selected in this case. I think this one is uh is color, green and based on uh on the options pass here. It will return a product which uh uh it's basically the same product, but only with the options uh available. So, for example, in this case, if we take a look I think it will only return yeah, it will only return option for sizes.

C

uh Next up, let's say the The Shopper selects uh size s, for example, and you will get back this time. If you don't find a query in your and you select the another option, let's say: I'm, also selecting a sizes.

C

And you'll get back this time. You will get a simple product.

C

You will not have any options left to select and you'll get a simple product with uh with a fixed price, and you won't get back like a price range like you will do for a complex. So basically, when the Shopper has select all the necessary options, the refine query will return a simple product. If not, it will return a complex product.

C

um Yeah. Also, uh like I said earlier, we were doing a. uh We have a Federated Gateway together with live search. So, for example, you can do uh product searches in catalog service.

C

This will internally use the live search to fetch the list of products, but you you can enrich the data return with the data coming from catalog service, so stuff, like uh attributes and uh options, are not normally supported by live search, but with catalog service you can actually get them. So, for example, if you, if you uh search for hoodie uh yeah, you can get back a list of uh these. All these products are coming from live search but, for example, the color and the and the size.

C

These options are actually coming from catalog service due to the Federated Gateway that we have. uh We have it implemented uh right now. We don't tougher all the all the features that are available in the in the monolith: uh graphql API, but with, uh for example, apms you can, uh you know, reach the data coming from catalog service to include stuff that we don't support. uh What one example will be, for example, uh this is an option.

C

This is a super simple example that I made using API mesh uh I've injected the information coming from a core uh Adobe Commerce graphql API to the catalog services like tier prices or entity, stuff that we don't support right now. You can actually uh get them from the the monolith that adobe Commerce API.

C

Oh yeah, so that's about it.

B

That's great thank you solution, so I think we have 15 minutes or more a little bit more just in case someone wants to ask questions or ideas to any of us, we'll be more than happy to to help you with.

A

Yeah I don't see question on the chat, but if anybody would like to ask live sure, please go ahead.

A

And you can always reach out to Sandra Lucian and rizwan later on, uh I think Sandra. You had the contact details and the slack Channel details on this slide. Yeah.

B

I'm gonna actually share it on the chat, so everybody have them easy access. Let me just drop this here.

A

Awesome great thank.

B

A

All right, thank you great now, we'll move to our second topic, which is community prioritization process and I realized I did not introduce myself. So those of you who don't know me, my name is parul Sinha and I work for Commerce Community engineering team. As a software engineer, so let me share my screen real, quick and then we'll talk about Community prioritization process.

A

So this is something that we kicked off last year, around October November time frame and ritesh samani uh spoke about it in Meet, Magento, New, York event in September, so he has this Dev blog post out that talks about this process. Basically, the goal for us to introduce this process was to empower the community. That was our real goal to empower in the sense that we want the community to tell us what they want us to work on, like you know where they want to focus our energy on.

A

So the second thing is that they can be the drivers of what goes into the code base of Magento open source. So this is really impactful and we really want to see Community getting involved, which we did see uh since we launched this program, and it's really simple like how it works is super simple. All you have to do is if you, you are a Magento technology Enthusiast, all you need to do is get onto GitHub login from your account make sure that you're logged in and then you have all the pull requests right.

A

So here you can take a look at all of these pull requests and the ones that you like, the most that ones that matters most to you. You can upvote the pull request, so you can just click on it, and here you have, um you know, reactions make sure to upload it. Now what we are doing at our end, let me go back to pull requests.

A

So what we do is we sort it on the basis of reaction and then yeah, and then we pick a set to work on right based on the maximum reactions, maximum upward thumbs up that the pull request has got, and we add a label called project, Community pit to the pr and then that PR shows up on this community dashboard. This is the dashboard that we have created for this process, and you would see that all these Community picked.

A

um Let me reload it once okay, uh you would see that all these uh label project Community picked labeled PRS. That would show up here. I need to there's something going on here. I feel because I'm not seeing all the tickets here today, I will have to look into it. What happened to the ones that were here in pending review and reviewing process I'll double check that, but that is how this whole thing works and then our internal team.

A

We pick, you know the ones that are ready for testing like here and then we start working on it. Make sure that you know um it works well. If everything is ready, if we have all the approval that is needed, then it would show up here in merge in progress. Then we open a Mainline PR and then that gets merged into the code base. So this is the whole process and from time to time uh we are sharing updates about.

A

uh You know the progress that we are making and that updates get shared under the dev blog on the Forum. So, like you would see I published January update two weeks ago, and this is the status that we have so far um for, like you know so far, we have processed two sets of PR the third one.

A

The third set is already on the dashboard, and um this is the status as of two weeks ago, and if you would like to see like the real time status, then you can always go back to this community dashboard and see so like you can see. We have already merged five PR's two got closed um and um one is ready to merge to win testing.

A

But this is the big chunk that you know uh that we also rely on the community to help us on which is the in review like right, like we have a team of community maintainers, they work on the these PRS. They review it once it's ready for testing, then our team takes it on. So we this is like you know.

A

This process is two-way, like we definitely need um to hear from the community, to kind of you know to work on the PRS that they want us to work on, but then we also need help from the community in reviewing them. So this is, like you know, a two-way process and real quick.

A

um Let me take a pause here and ask if there's any question around this process, anybody.

A

I think there's something in the chat: okay, Sandra has to leave. That's fine. If not, then I just wanted to uh quickly.

A

uh You know give a plug about our newsletters, so I've been publishing monthly newsletters and if you are not subscribed to it, if you're not getting, um you know notification about the newsletters, then please make sure that on the Forum here you know when you go to news and announcements you should see here subscribe for me, it's unsubscribed because I'm already subscribed, but if you're not make sure that you subscribe to this news and announcement so that whenever the newsletter is out, you get an update.

A

The reason I'm stressing on this newsletter is because you know we covered updates from both Adobe Commerce and Magento open source, and always there would be a section that would talk about Community prioritization process. uh Like you know, the progress that we are making so I'll make sure to have this section all the time on the newsletter. So this is another place that you know if you want to see how it's going. This is a good place to go for.

A

That's all I have. If there's any question I would be happy to take.

A

Okay, if not, then um I think we can wrap it up thanks for joining everyone and we'll meet again in the next hangout session. Thank you.

C

C