Red Hat OpenShift Developer Experience Office Hours, 2 Mar 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Developer Experience Office Hours: New serverless features in OpenShift 4.7 Console & What’s Next

Description

Join OpenShift's Developer Experience experts for our regularly scheduled program filled with cloud native, Kubernetes, and OpenShift tips and tricks for developers.

A

Good morning good afternoon good evening and welcome to another edition of the developer experience office hours, I am chris short executive producer of openshift tv and today I am joined by three of my fellow red hatters uh serena from the future nichols uh brian jarvanin and jay. I won't try and say your last name, but please uh start the introductions.

A

uh Jay go ahead.

B

Yeah, hello, everyone, it's really nice to be here. uh My full name is javertin. You can call me jay and I am working with redhat as a ui engineer in open soft. Basically, and so it's it's pleasure to be we're.

A

Happy to have you brian.

C

Yeah, I'm a developer advocate in red hat's, open shift team um I'll, be keeping an eye on questions in chat and also, if you have future topics for this, show that you'd like to see us address. uh Definitely let us know in chat or otherwise and we'd be happy to try to cover those topics as well or q. A you have regarding this specific uh episode, um I'll be helping wrangle, those in chat. So uh definitely let us know if you have questions about anything we're covering today.

A

C

You ryan and serena.

D

Hey everybody serena nichols, I'm one of the developer, tooling uh pms, focusing on openshift developer experience in the console, and uh so we're glad you guys are all here and jay is one of our epic owners, specifically while very frequently working on serverless features inside of that console. So we're really happy to have him here and my dogs all say hello as well. So.

A

I was gonna say.

D

A

Serena has some dog duty to take care of today. So yes, if serena disappears for any reason, that's because she's chasing a new dog, so yeah.

A

D

Maybe you could.

A

Start yeah, you wanna, you wanna, kick everything off here: buddy.

E

B

I hope you guys can see my screen.

A

Yes, I'll get there.

B

So sir, would like to start.

D

Yeah sure so yeah today, what we're gonna do is we're gonna demo. um Some of the four seven features that we've added into the console um and then after that, if we have time we're even gonna start demoing, some of the things that we're looking at that we've already completed development on that will be coming in post, four seven, so some really exciting stuff here.

D

um So I think jay's got a couple of slides here, just to again remind us on what cyrillus is as well as on the next side um introducing keaney. Oh, I.

A

Forgot, I forgot to say, remember folks: some of this stuff is future stuff. Some of this stuff is now stuff. So keep that in mind.

D

Right yeah thanks for saying that no.

A

D

Yeah and so we're going to be um talking about k native and then specifically with the console when the openshift serverless operator is installed, we have the ability to get serverless features as well as eventing features, um so this first one. This first feature that was added in 47, which is now available, um we'll be able to see some things in the admin perspective so I'll just hand over to jay to to demo what those features are.

B

Yeah, so I'd like to start with the first one, so this is the cluster of oversight cluster and, as you can see, I have the operators installed, that is redhat openshift, serverless 1.13, which is needed, and I also have red authentication camel k which I'll be talking a bit later about. So the first feature goes here is we are now providing option for uh admin as well? In like admin perspective, as you can see, we have a serverless nav section, which has both the separate things like serving and eventing.

B

If you click on that, it will list you all the services, revisions and routes, and likewise we have for eventing, which lists all the event sources, and it has all the features like filtering or searching the brokers, the triggers channels subscription, and we even have ability to create. From this I mean you can go ahead and create any of these save for event sources.

B

So that was about the first feature we have. So I think we should talk about the second one. Now.

D

Yeah jay, I was also going to say you want to just remind people um if you go back into the developer perspective, that they can search for those elements and then add them to the navigation, um because it's always a nice, a nice thing to give so yeah. You need to go to dev yeah.

B

So that's a great.

E

Feature because.

B

Many times when we are working, we would like to see something. For instance, I want to have my brokers handy, so I can just select on broker and say, add to navigation, so this will be preserved here at any point in time we can come back and we can see the list of brokers available.

B

So that was about this, so you can have a number of navigations here.

D

Thanks jay for showing that too, even though that's a an older feature, I think it's something that um when people realize that it's there it's pretty exciting for them. So it's a good thing to show.

D

uh So the next piece are you gonna, look at the event source catalog.

B

Yes, so if you go to develop perspective, we have option here under add you can see the cards over here. You can go to either through catalog or event sources directly. So, if you click on that, it lists all the sources.

B

So the objective behind is to improve the user experience by providing scalability, because even sources are growing and we also support camel connectors. We call it camelots over here, so you can see like jira salesforce telegram slack. All of these are like powered by.

D

Nice and so jay, can you just go over? I don't remember specifically what happens as far as operators are concerned. For this is there a specific operator we need to have installed in order to expose those.

B

Sorry yeah so again, this is the piece so.

D

Yeah initially.

B

I spoke about the operator which I have installed, so that is red hat integration, camel k, 1.2.1 powered by red hat. So as a user, we need to have this operator installed and we can also create integration platform. So, whichever namespace you create this, you can see all the cameras over there or camel connectors over there listed, and you can also access to that. From eventing perspective, I mean just by clicking on even sources with one click.

D

And again, just to just to kind of talk through this, if the camel, um if that camel operator was not installed, you'd still have event sources, but we'd be limited to what is it four or five of them.

B

Yes, like api server, source uh container source ping source, sync binding and kafka source, if you have created the cr for that which is provided by serverless operator itself, just to highlight so you can see here kinetic kafka. So if you create click on this and if you enable it the source, you can see it so.

D

I see okay, excellent thanks for walking through that.

D

um And I think also just to kind of note the event source catalog. We I think we mentioned some of this in previous um office hours, but we have in 4 7 we've changed kind of, though the experience of our catalogs a little bit trying to make things more consistent throughout the sub catalog. So you could still get when you go to the developer. Catalog you'll see event sources inside the larger developer, catalog or you can directly go into the event source, catalog and just the event sources.

D

So the next piece you're going to talk about channels and subscriptions.

B

Yes, so uh I think already in previous sessions or a couple of times, we have seen about the different ways, uh events uh or basically can be consumed uh like one way is direct connection which is like source to service, and in today I mean in this session, we'll be talking about channels and subscriptions. So that's a way where we can have multiple subscribers for different events coming through different sources. So, as you can see in this particular diagram or this particular visualization, so let me take you to the topology view what we call it.

B

So this is the sample project which I have created.

B

So if you guys can see over here, uh I have a channel which is in memory channel, and I have two of the sources uh which are being synced to that one is ping source, another one is camera source, and then we have these two services which are receiving all the events coming through all of these sources.

B

So in channel we don't have mechanism for filter. So all the events emitted by these sources will be subscribed by these services over here. So this generative service, which I created, which basically predict the locks emitted. So if you see over here, you can see hello channels. This is something uh being emitted through ping source and I would like to add one thing to that: you can do all of these creations through gui or through web console as well.

B

So, for instance, if I want to show you guys I'll go ahead and click on channel, so you can see there is options. uh If you would have enabled kafka channel, you can even see that so I'll try to create a channel now, so we have our channel now. Next, we need is uh event source, so I'll go to ad flow and which we have seen I'll click on event, sources, I'll, say ping source and if you click on this create this is the new thing which we added in 4.7.

B

We have the ability for user to switch between form and yaml section. So earlier that was not the case, so user is no more limited and all the data whatever they are typing here- will be preserved here. So I will say per second and sync: it to a memory channel. I will not have the app name and I'll say create, so what this will do? It will sync the ping source to in-memory channel and which can be visualized in topology as well, and let me quickly go ahead and create one more event source.

B

So this time I will go with api server source. Yes, so already I have created a customer bindings to watch the events and a service account associated so I'll. Just select the survey service account here we are calling it events essay and there are different modes as I'm going to watch events I'll call it resource and api version for events will be v1 and kind will be event.

B

And if we click on say.

B

Create you can see uh both the sources are synced to in memory channel now. Next, we need is a service which we need to subscribe to, so I'll go ahead and I'll use container image, and this one the image which I use always frequently.

B

So it's more like a logger service which uh prints the information going over there, so I'll call it channel display, because that is what we are going to use or showcase I'll call it zero, and we have to select creative service of the option in resources because we are going to create a serverless application and I'll click on.

B

B

So we have one of the service ready, uh let it come up and let me go ahead and create the other service just to showcase that so I'll go ahead to container image again I'll use the same image which I used last time and I'll call it channel display one.

B

And I'll select again kinetic service, as you can see, all of these operations have been performed through web console, uh no need to go to cli. If you don't like it much and if you do feel free to so, as you can see over here we have our services up. So let me try to rearrange it.

B

D

So jenna just type in here a little bit too jay, because the other thing that's kind of cool in 4 7, is that topology layout now persists too. So, as you're rearranging things the next time you come back into this project, you don't lose that which is awesome. um So I just want to mention that as well.

B

That's a very good point and I think it will use so for user to persist the layout. So now I'm going to create a subscription. So I'm just it's easy. This is I just dragged and dropped on a service and you can see that it is asking you for the name. If you want you can modify it. I won't for the session yeah and I'll use another one to connect to another service.

B

So I have my memory channel being subscribed by two of the services and soon you will see the container logs emitting the events or showing the events whatever is being subscribed or by this services.

B

So meanwhile, let me go ahead and show you that, yes, as you can see, all of these cloud events are being provided here or you can watch for so based on this, the service could be anything any real use case where you want to perform some action based on some cloud events.

B

So that is what what I wanted to showcase here and, as I told earlier, the with channel you, we can't do filter, basically, so all the events emitted by this api server source, where we are watching for events or ping source which is emitting on this some time interval. So all of these events are being sent by both of these services, so that was about channels and subscriptions, so uh anything to add serena or shall we go back to the slides.

D

No, I think that's great you're going to go over um brokers and triggers next too right.

B

Yes, so in this slide, maybe.

D

B

Visualization demonstrate that I mean if it makes more sense there yeah, so next is brokers and triggers so the drawback or or our value, the one of the benefit which we have with bro current triggers is like. We have the ability to filter it out, so we can filter any of the cloud events based on uh whatever we want to and uh or attributes or any of these things which were not possible with channel. So this is how the visualization would look like.

B

So let me take you back again to the project which I already created.

B

So if you see over here, so this is a broker and we have three of the event sources which are being uh synced to this broker, and then we have again two services and if you notice this connector, we have this one perpendicular line over here. So this denotes the filter. If you click on this, you can see the filter. What is the filter I have applied over here, like it's def dot, kinetic dot sources, dot pain. So I wanted this service to only get basically events from ping source.

B

So that's what it does and for the other one I don't have any filter, so it will still get all the events from all of different sources. So let me quickly show how we can create this through gui again so I'll, go to admin perspective this time for change, and I will go to eventing I'll select the project which I was surfacing earlier and create on broker.

B

So this is what a simple broker would look like I'll call. It default.

B

And I'll hit create, so we have our broker created. If we go back to the topology view again, we should be able to see even broker over here, so we have our broker now. Next, what we'll try to do we'll try to create a ping source again or better? I can let me create one so I'll repeat the process. What I did last time so I'll go ahead and create on pink source I'll, say ping broker.

B

So we are thinking of the broker, but still you have options to choose any of these things, so I will continue with broker and I'll call it ping source.

E

B

And schedule it to give I'll give per second so I'll say create. So now I have my ping source ready.

B

And just to show you guys, I mean uh we can even point it to some other addressable object here with easy, drag and drop. If you want to, you can point it to imc or you can add it to a broker itself, so I'll go ahead and create another event, source that is api server source.

B

So again, I'll use the same thing to watch for events.

B

And I'll go for resource as I'm watching for events I'll select event, essay, a service account which I already have created and next thing is say, hit create.

B

So, as you can see over here, uh as I pointed my ping source one to uh memory channel, so it has pointed to that, so let me bring.

B

Let me make a change and this time make this pink source point to broker. So we did that and yeah it has updated.

B

Yes, so next is again service, so quickly I'll, go ahead and repeat the process which I did last time just so that you can visualize what is the difference and I'm going to create again a connective service, I'll call it broker display.

B

B

And I'll quickly create one.

B

B

Call it broker display one.

B

And select creative.

B

B

Now we have all of these things over here. We have two of the revisions and we have our broker.

B

So let me make it set to the screen and keep the services in one side. Yes and.

E

Sorry yeah, we have so many.

B

Things on top, I don't want these options. Yes, so next, what I wanted to show you is like how to create a trigger, so we have our event sources which is being sent to a broker and with easy questions. Again. If you click on this and drop it to a service, you get option to add a trigger.

B

So, as you can see, this is the trigger name and we can provide the attribute here. I'll call it type and value will be dev, dot sources, dot, ping. I wanted only ping source to be received by this service, so that's more like a filter here, I'll say, add and now I'll drag and drop another trigger to a different service. So I don't want any filter over here. So I'll do the add again and as you can see so we have everything set up.

B

You have to wait for some time and slowly we'll see the spots coming up or the moment events are being received and we can see in the log the filter happening only for this particular service, not for the other one. So.

D

Jay, I have a question for you: yeah. um If the user didn't want to do, drag and drop, is their ability to do like the right click course or go to the actions menu for the broker.

B

Yes, okay, so if you don't want to do, drag and drop, you can do a right, click and say add trigger and you will get the same option like if you want to. If I want to subscribe it to a different service. Now I can do that right away, okay and the same thing and then achieved with admin perspective as well, so just to highlight that if you go to eventing, if you go to brokers, we have option to add trigger here as well.

D

Okay and then that would also be the same menu like if you're in in topology, and you have your um broker selected. There's an actions menu on that right hand, side panel, that that would also allow you to do it from there as well. Right.

B

Yes, so if you click on any of these, I think you can see now the parts are up. So if I click on these logs, I should be able to see only the logs from the ping source that is test hello, and if you go back to the other service, you should be able to see the logs from both the human sources.

B

It's gearing up, I believe, yeah.

B

So, and if you click on this broker, you can see the details of the subscribers like okay, these are the services and you can click on so filter. It will show the filter over here and yeah likewise for channels as well. You can see all of these details in the sidebar inside navigation and this action menu helps you to perform the same object things whatever we can on the right click on the context menu.

B

So that was about the brokenness, which I wanted to so yeah.

D

That's awesome, let me ask you a couple other questions. If you don't mind, um could you go back to the ad page.

E

D

So, on the ad page, we have the ability, we have a tile to create channels and we have, um but we currently we don't have the cr the ability to create brokers from the ad page, correct.

B

B

So that's the feature is provided in the admin site, but yeah. I think uh this is something we will be working on. I think we'll be having it soon in ad flow as well. Yeah.

D

Yeah, that's what I was going to also say because it so you had you already added your broker's uh navigation item. So if you had that there you could just go down on the even in the dev side right and then you could use that create broker up there on the top right.

E

D

E

D

If you wanted to create it from there- okay, okay, great- I just want to let people know that that still is you can still create it from the ui. But it's just yeah um yaml forms right now. So one.

A

D

That we've done enough a lot of work on and the last couple of releases and continue to do is trying to have form-driven experiences in addition to yaml driven experiences and we're just not there in all the spots. Yet so.

D

Okay, and would you highlight uh as to why you would want to use the the brokers over the triggers? I'm sorry, the brokers over the channels.

B

Yeah, so so uh the main benefit I see of using broker over channels is the ability to filter it out, so assume that you have n number of event, sources b from slack telegram, ping source, api, server or kafka, but- and we have like three or four subscribers or different services listening to those, but as a user, or we have a use case where we only want a particular service to receive events only from sake, after not from other sources, so broker helps us with that to achieve that filtration logic, with the ability to filter it out so, which is something like providing different attributes like one such thing.

B

I did here in this particular trigger I added type for ping source. So what this means is this particular service which only receive events being emitted from ping source, not other sources which are being synced to this particular broker. So if you say add trigger, you have option to add multiple attributes over here.

D

Awesome: okay, thank you, um chris orion. Is there anything in the chat that you guys want to bring up around what jay's demoed for four seven so far, yeah there's.

C

A

C

Yeah, oh I'll, take it all right.

A

C

There's a couple questions that popped up in chat, one that I'm gonna but in front of the line and ask from the um from the ad view you you had I'm going to throw my own question in so from that when you did add, um there was catalog event, source and then channel. Is there a way to filter these to to get um just k native things or just uh I've? Seen in the past, like a check box on the side of things that are only operators?

C

Do we have anything like that uh to help filter um what we're looking at.

D

So, just on the ad page itself, right now it looks like a tile view. We are just so you know in the future. We are also going to redesign this, so things are kind of more aligned with everything, that's k-native-based or pipeline based, etc. But I think jay was just going to go into the catalog itself. When you go into the larger developer catalog, we do still have these subtypes. On the left hand, side the differences in 4 seven. Well, I'm sorry! Let me step back pre, four. Seven.

D

We used to have check boxes there that allowed you to see multiple types at once now, when you go to developer catalog, you see all of them. If you want to drill in you only see a single type got it if that makes sense, um and when you're in the larger developer catalog, you just do still see those labels on each of the tiles that are indicating. Is it a builder image or a helm, chart uh event, source, etc and.

C

There's event sources there: okay great, and um there was also, let's see if I scroll back and chat a little bit. I saw openpixel had a question about: what's the difference between k-native kafka and amq streams,.

B

So uh the difference between k, native kafka and mq streams is a native. Kafka will provide us with the kafka channel and the kafka source, but in turn, mq streams will help us to have the kafka cluster up, create the bootstrap server, top copper topics etc, so that kafka source can consume. That particular thing so that.

C

Helps you yeah that uh open pixel? Definitely let us know if that doesn't cover that particular question. um We had another one from open, pixel um saying that uh triggers are a basic kubernetes resource uh correct or is that a k-native resource.

B

So the triggers which I sowed- that's a genitive resource. It's kinetic triggers and I think the question is: is there a way to uh trigger events from external source running outside of openstack, say database service on cloud.

B

So again I mean uh to answer that uh yeah. I think if it is a if it is a public url which can be accessed, I'm not sure, but I think there is source for particularly that just like we have for different sources, be it twitter or any other thing. So I think that can be consumed.

B

So I'm not very sure with that answer like how it can be done if it is something external but yeah. If we have external service which is addressable object, then that can be synced to in openshift.

B

So what I mean is if you go to uh developer perspective again and if you go to any of the event sources or any of the scenarios, and we have the ability to sync url. So if you notice over here, so it takes any valid uri. So this will be useful if we have any addressable entity, even if it is outside of question.

C

Awesome um so openpixel was asking a clarifying here uh so he's uh imagine our r k native streams equal to kafka and then k native kafka equivalent to uh the integration into openshift. That doesn't sound quite right to me.

C

I'm not sure about uh how to.

D

Answer that one.

C

D

I'm not sure about that one either. We might want to get more clarifying.

C

D

C

Yeah yeah we'd need to, I think, have folks.

B

C

That team in there.

B

Yes, I think thomas answered there, so basically m2 streams is red: hat separate version of stream g kappa it's in that sample string kafka is a community operator. So if I take you guys to the admin perspective again and in the operator hub, uh you can see, I think stmg will be something we'll be aware of.

B

You can see this community version so likewise we have mp streams.

B

Which is uh like production version from red hat and if you wanted to try out kafka source, so let me show you right away. Let me go to kinetic eventing, namespace and I'll try to create kinetic software.

B

So if you click on create, you should see the option like channel if you want to enable it or source so I'll, enable it and channel as well I'll say enable it. I want kafka channel to be there now as the default channel and, if I hit create, so it will bring in those crds uh basically over here, so it will take a bit to uh instantiate and after that under event sources. We should be able to see kafka source as well as under channels. We should be able to see kafka channel.

B

So let me go to sorry developer perspective.

B

And, but still with this, uh we still need to have empty streams or stream g uh in order to have uh bootstrap server or you need to power or software topics over here. So if I go to developer perspective now under add flow, let me select yeah.

A

B

Space where I was in, if I select say, event sources uh and we can see the kafka source over here, which was not there earlier, so that was the cr which we created and if you click on this, this is what I mean bootstrap server. So once we have a stream g, we need to have a cluster setup, so this will provide us these urls or the topics which we want to subscribe to or listen to. Basically,.

E

D

Thanks for going.

E

D

That and I think ben just answered one of the questions as well right um in chat.

A

Yeah. Thank you. Thank you. Thank you. Thank you. Thank you.

D

So much in chat.

A

Right now, by the way, yeah and q streams helps you deploy and manage coffee clusters inside your openshift cluster, okay, nested clustering, good stuff.

D

You know, uh and and a few of these like as jay, has done as showed as as soon as some of these operators are installed. We expose additional elements inside of the ui so that you can um can access the features for them, which is cool.

D

Okay, so jay, I wonder, is it time for us to flip over to the future?

D

Okay, so um the first thing we're going to talk about or show, I guess is one of the things that, um if you're creating a serverless function, uh it is visualized inside of the topology view, even though we don't currently have the way to create one inside of the console right. So there's the ability to create them through the cli, the k-native cli, and I don't know if you were going to go over that or you're just going to show the visualization jay.

B

Yeah, so uh let me show you the visualization first, and maybe I can talk about the steps if someone wants to try it out. So sorry, not this one upcoming demo. So if you see over here, we have two creative service, so one is powered by serverless functions. That's why we have this effects icon and a different background color. So this is the way uh you can identify anything escalated through serverless function in topology, and if you wanted to create it, we can go to cli and.

E

B

So let me create a new project, I'll call it just one two.

B

So it's very simple to create you can just save pay and fun, create uh note test, asymptote I'll be using, so it basically creates a projector node test. If you go to that and do for dot, it will open a vsphere for me and just to showcase what it does. It creates our index.js and we can see that we have a get and post where it returns, the query param and the payload. This is the simple thing it is trying to do so.

B

Let me go back to the cli again and I'll do 10 deploy so basically what it does. It tries to build this and deploy it into any of the registry. So I'm using docker here.

E

B

So it is trying to build that thing and deploy it into the particular namespace. So it will take a while to do that and let me meanwhile go back to the slides and I'll revisit this and show you guys in that name of space concepts up so.

D

The other thing we could do j, if you don't mind if you go back to topology view and just zoom in a little bit, so people can see just a little bit uh closer.

D

The difference between that box on the left, so the background color for a canadian uh first serverless function is, is like a light purple and then to the left of the k native service badge we're showing that fx icon. Like we mentioned this is not in four seven. This is not released, yet it's in the future, so that icon will likely change before we get this in the product. But it will be something similar to that to denote that it is a serverless function.

D

And then, if you look at the item on the box on the right, that's just our regular k-native service, where it's a gray background visualized. Similarly, but it has a gray background and has the k-native logo, and then I think, we've also talked about, but we haven't figured out exactly what we're doing. If you select one of these items in the side panel jay, we talked about somehow probably in the details, tab there denoting that it's a serverless function as well right, but we haven't. We haven't gotten to that. Yet.

B

Yeah um and just to add being painted service, it has all the ability of a serverless security service like traffic spreading. Everything can be achieved.

E

B

If I go back to my function is deployed, so let me quickly validate this. If I go to uh test.

B

As you can see, uh it is up, and if I click on this particular url, it says query I'll, say.

E

Name, let's do j.

E

And let's do soft.

B

So it just does what I saw in the code, so it could be anything.

B

I'll go back to the slide again.

D

Excellent, thank you, okay. So the next piece is we've added a couple of extra advanced scaling options for the k native service. So um are we able to show that in an edit, as well as in a create.js.

E

D

So maybe we just demo the edit piece, maybe.

B

Yep make sense, so let me go to yeah back to the service functions which I created. So if you do right click and say edit note test, it will take you back to the edit flow and you have this option over here like scaling. So if you click on that, we can see a lot of options over here uh talking about auto scale, window sentence, utilization and all of these pieces.

B

So as we are editing it, so it is not populating the default values because nothing was provided in the beginning by user, but if you are going to create a new one, so just to showcase uh in container image, we even populate the default value, what they are currently.

B

So, if I select native service now- and I click on scaling- we can see that 60 seconds is the auto scale window. So you can change it to minutes r. It goes till one hour, so these options will configure. You can set up contingency utilization, limit, target and min and max part which we already had. So all of these capabilities are there.

B

Actually, I think.

D

I was just going to say now that that's great, so now we have the ability to do that upon creation or even after on an edit if you want to make those changes, and that gives us some consistency with the support level in the cli now which we didn't have in four six right.

D

um I should we didn't mention here like what version of the uh openshift serverless are we operating.

B

D

B

So we are currently into 1.13 of opencv serverless awesome. Okay,.

D

Great and then I think, there's one other feature that um you guys have been working on on the dev side as well right around keynatifying, something as some of us like to call it um so again, just to make sure people know this is not four seven. This is going to be something in the future.

D

Maybe you can talk about that. One as well.

B

Yes, so the purpose or idea behind this is assume that I have a workload and I'm new to canadian world, and I wanted to try it out with my workload, be deployment or deployment conflict, so I'll quickly go ahead and create a deployment, I'll use, basic, simple, hello, opposite image, and if you click on create, it basically creates the deployment for us. So once the deployment is running, uh we'll see an option to migrate it to a creative service.

B

So let me take you back to the slide just to demonstrate. um If you talk in the yaml thing, it's something like this. So when we create a deployment, we have to create service route. If you want big space, a lot of those things so with creative service, it's just small, even with uh the yaml I mean we see a lot of how to quality. I mean reducing the specs in the aml, but now with a web console with the easy click you should be able to achieve that.

B

So uh I think it's taking some time to come up, yeah continuous, creating! So if you click on this, we can see create connective service over here, so this option is shown across in all the context, menu or action items. If you click on this, it takes us to the new view where it shows that okay, this is the image we are trying to use, which was part of the deployment, and we have the ability to again tweak any of the advanced options.

B

Whatever is being shown over here, be it health checks, scaling, resource limit, etc, or you can just choose to ignore it. So, with one simple click, we should be able to have our creative service out of the deployment as well, which user can try it out. So this was the deployment and we are not building a deployment at this stage, so it will take a minute. I guess a few seconds not a minute. Sorry, and yes, it is up, even if it takes a few minutes. That's pretty amazing.

E

B

Well, and if you see what open shift is it's nothing but hello concept? So let me quickly, sorry, let me quickly go and show you generative service.

B

Yes, it's up and running, and it has all the ability of alternative service, perfect splitting scaling and all of the features which it offers.

D

So yeah, this is awesome, so just to mention too, like we said this is future right now we're calling it create k native service, we're going to be doing some, I think, validation, if that's really the name, so things might change by the time that comes out, um but this is definitely a feature that people have asked like. How are we? How are we able to take an existing service and make it be serverless?

D

So I think this is like really interesting work that the dev team is doing and and has started off on, so it's pretty cool thanks for sharing all this jay looks really good and again in chat like if we would love to hear any any of the comments and what you've seen in four seven um as well as some of the works that we're doing post four seven, please, you know, let us know what you think and jay. I think yet another slide there.

B

uh I think we are done with the slides.

D

Oh okay, okay, awesome, awesome! So is there any other questions in chat? um Oh.

A

I'm sure there is, let me yeah real, quick.

A

Alrighty, can you see the nodes view in four seven? I have an issue with the formatting of the table in my cluster on four seven: zero, sometimes loads correctly. Sometimes the rows get misplaced ooh. Just in node's view, though, that's.

D

Weird interesting, I think jay you're running like a homegrown, not homegrown, but.

A

You're you're running.

D

Like a four post, four seven version, so um you could still try, but I'm not sure if it's gonna look exactly like what people would see on four seven.

A

uh Be kind asks what is the roadmap for openshift serverless functions, component services and eventing are mature, but what about designing and deploying functions like in openwhisk or other? You know functions as a service type thing.

D

Yeah, I think we need to pull in karina or nina um the pms that are aligned with openshift serverless itself. I don't have that uh information. I'm not sure if any other red hatters that are on chat might have that.

A

D

A

D

Have I don't have that right off the top of my head.

A

Yeah, I don't either so me. Oh you, oh ryan blah. He did it. Okay,.

C

Yeah yeah, I'm not totally sure on that one. I know we used to have a way of running openwhisk directly. I think I'm not sure if there was an operator or a template or something, but um if you search for openshift and open whisk, you might still find examples.

C

um I think that was back in the openshift 3 era, though, and a lot of our functionality for serverless is now basically covered by k native from my understanding, I'm a 100 certain on that and I'd be happy to defer to other folks but um yeah. I think, k native and camel k is kind of the future uh for the kind of cloud native community. That's building directly on kubernetes.

C

I don't think openwhisk had anything um specific to kubernetes like it wasn't built out of crds or or I don't think it ran on the control plane um in the same way that k-native does um so k-native gives you a little bit tighter integration and more um more community support around kubernetes, focusing on kubernetes.

A

I'm answering a wrong a long question in chat, so let's just talk about it are the features demonstrated part of open source knativ and the console and developer experience is open. Shift value. Add question mark I I'm thinking they're asking like: where does open source start and uh open shift began? You know.

D

They might yeah, and they also might be asking specifically on that create k native service, because that's not something that's part of and jay you can. Maybe you could also explain like the implementation of that create k-native service. um That's not something! That's an api! That's create that was provided by k by the k-native.

E

D

Community correct.

B

Yes, so as part of create creative service, it is something which we have handled completely into the gui in the first version, uh just to get the feel of it like how it works.

B

So in order to spain, uh basically, what we are trying to do is get the image whatever is being used from the internal history and try to map the specs uh to the connective service. So, for instance, if you have hpa, you will have a min scale max scale and other things are like the beauty, involvement variables or whatever is there in the container?

B

Basically, the word spread, so everything can be used in serverless and we are providing the option as well to user more like to validate what is getting in, because in the form they can see. Okay, what is the image? What are the advanced options being used, so they can tweak with it and just create it, and but there is plan to have it uh completely driven by the apis.

B

So I think that will be something to have more consistent and in order to support even failovers or graceful rollover. I'll call it in future.

E

A

Chat is pretty active, but uh paul is in chat. Answering lots of questions so yeah odd man. Support is either just mers or about to merge by the way, be kind.

A

Yeah good call serena on that one um be kind you.

C

Were asking uh be kind as a question that says how are these run times being packaged um and I'm not entirely clear on what they're speaking to as far as run times, um I know for my code, I've been using still source to image, to compile some of the application code and then just deploy it as a k-native service.

C

So it can be pretty simple from the ui, but I'm not entirely sure. If that's what you're asking about in terms of runtimes yeah, like which one like so many.

A

C

A

C

My perspective run time is which version of node.js are you using? Are you on the node, v12 or node v14 that'd, be my application runtime, um but there's a lot of ways to define runtime.

A

C

A

Functions are using cloud native.

B

A

If that answers the runtimes question.

B

And in several functions we have option like you can pass with parameter like hyphen l, as you can see, I mean it can take like different times if it helps.

B

C

Oh good uh yeah, that's uh good feedback. um Cheddar mint and uh bee browning are chiming in and chat, saying that serverless functions are using cloud native, build packs um and that's a good point. So we had a source to image build strategy. A lot of that build process is getting moved into what was k-native build and is now techton some of that functionality used to be covered by k-native and has been migrated into the tecton feature set.

C

So that's usually how the the build would be run for kind of packaging, up your your runtime services and making them available on the cluster.

A

Yeah sounds like we ought to have someone chat about build packs on here, sometime for sure, um and it's buildpacks.io is the website. If you want to go to, they have a demo like right there off. You go kind of deal.

A

And also to mention, it is also a cncf uh project. I forget, if it's incubating is what it says so yeah.

A

I will look, do you know any idea?

A

Yes, I do jp dave hit me up with email style.

A

Great chats yeah lots of people.

F

Yeah I'm just thank.

C

You scrolling back.

A

So yeah yeah cool.

D

I was wondering ryan: do we want to do we want to put that survey in that we had um in the last couple of weeks. I think we.

F

D

It for a few weeks right.

F

Sure yeah be happy to post that if you have.

D

F

It and if not, I've got.

D

It right yeah, okay,.

F

D

Got it yeah? Maybe you can just quickly uh we're asking people who haven't already filled it out to fill it out um to indicate what they like this developer experience office hour to to kind of um cover. I guess right.

C

Yeah we had a lot of great.

D

Suggestions today.

C

I know there was one on: uh let's see what was it called uh fungi funky, maybe yeah.

A

um Was that linked to.

C

Do we link to that I'm going to scroll back in the chat and see if I can find it yeah, plus one to corkus fungi monkey and openshift server list demo? Okay,.

A

Yeah, like always, go to learn.uh openshift.com, um and if you wanna this, actually I think you can install the serverless operator in our sandbox. Now too, um your your mileage may vary. It will probably be pretty slow, uh I'm not sure if it'll actually like fully work, but you can definitely get started with crc on your local machine. At the very least.

D

Cool yeah, the developer sandbox today um pretty sure there's no server list on that. Yet there.

A

Is no okay, I'm not sure whether, like that project.

D

I could be wrong, but.

A

I don't think yeah.

D

C

Yeah I just I just.

D

C

A

C

In there, okay, cool.

A

But yeah definitely the the learning channel and you can definitely run this stuff on crc seriously for sure yeah, yep, yeah and crc.

D

Stands for ready containers.

A

Sorry, I'm using vernacular that people might not know shocking.

D

I guess we could also point to the red hat developer page that oh.

A

F

Guess maybe you've.

D

Already done that, okay, perfect.

A

D

A

D

A

And, like I said, there's.

D

A

Two k native books we offer for free. I will redrop those in chat, k, native cookbook and k native patterns.

A

um I need to make that k native patterns description, much shorter, but yes, they are both available for free on the developers.redhat.com site, just head over to the ebook section, if you're navigating along and there's a link directly to it. If you need it um great awesome, this has been great. It's a really deep dive. I I am now intrigued by k-native, more than I was before so good job. On that one jay, um do you still need 16 gigs of free memory to run crc?

A

No, they are actually working on that we're trying that's at 12 right now. I think it's going. The next step down is nine right, ryan, yeah.

C

Not sure I don't.

A

Know I I got a.

C

Super 32 gig machine just so I can run crc, which I know is not realistic um for for the general population, um but you the sandbox, definitely does something to kind of help make those environments available.

C

There's definitely limitations on that you're, not admin, and you can't load all your operators, but you know that's part of the part of the trade-off there. um Crc does seem to be uh pretty active. I think they're just about to do a 1.23 release which um I may have the newest version of openshift in there cool I'm keeping my eye out for that. I saw a beta release, but I don't think it's got an official um we've got it on our internal ci servers, but not not out for the public. Yet.

A

Yeah I have the what is it called like cloud envy or whatever, where it's like? I can. I have a server in my basement and I run stuff on that. So I have that luxury.

A

um So yeah, if you can get your you can yeah, you can definitely get your hands dirty on this with crc and definitely look at the candidate of cookbook. That jai is showing or jay is showing god I did it again. Jay is showing on the screen um and if you have any questions as always feel free to reach out, I am c short at redhat.com.

A

uh Chrisshort on twitter feel free to hit me up anytime. I can connect you with any of the folks on this call or you know otherwise, uh like I said, uh jp dave has uh idm needs, so I'm he has my email and knows where to find me. So I'll get him hooked up in the right place and if you have questions I can do the same for you.

A

um So, thank you, everybody for showing up today and walking through this great job jay, as always, um wonderful demonstration, very good, deep dive, and thank you all for joining out there and asking your questions. We wouldn't do this without you. So thank you very much and uh just a quick heads up on.

A

What's next on the channel today at 1 eastern 1900 central european time, uh red hat, advanced cluster management presents cluster pool scaling and automated dev clusters, which, if you have to manage dev environments, this is going to be awesome, for you so definitely check that out again, that'll be in about an hour, so yeah stick around or click that uh subscribe and follow button wherever you may be with that said, uh we can safely say goodbye. If you have further questions again, reach out, don't be afraid and thank you. Everybody stay safe out.

A

There stay healthy and we'll catch you next time.

C

E

E