Red Hat OpenShift OpenShift Commons Briefings hosted by Karena Angell (Red Hat), 27 Oct 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: OCB: 3scale API Management Overview and Operator Demo

Description

Join us for a quick introduction of 3scale API Management, a deeper look at the 3scale operator application capabilities and a short demo of the latter.
Speakers: Satya Jayanti, Miguel Soriano, Eguzki Astiz Lezaun (Red Hat)
Host: Karena Angell

A

Hello, everyone welcome to another openshift commons briefing, I'm karina angel and I'm an openshift product manager, and I'm really excited to be here today with the three scale: api management team and really waiting for this one. So I will let them introduce themselves and really excited to hear about all their capabilities. Take it.

B

Away hi hi everyone uh and thank you to the openshift commons team for giving us an opportunity to present uh my name is satya jayanthi, I'm the product, marketing manager for api management, at red hat and with me, are my colleagues from three scale: team, uh meg, well and iguitsuki uh miguel. You want to introduce yourself quickly.

C

Yeah, thank you. Satya hi everyone. My name is miguel soriano and I work at red, hat's three scale team on the operator's soup team.

B

If you can introduce yourself as well.

B

On I think, you're on mute.

D

Sorry you're you're right. My name is woozky astith um working together with miguel and thomas in the so-called operator team in three scale, and developing fixing backs of a three skill operator together as well with um apicas operator.

D

Both operators of the three skill team.

B

Thanks, so what we want to do here is we have a short agenda for uh today. uh What we wanted to do was uh tell you uh api management in the context of the red hat integration portfolio, which is where the three scale api management product and the offering belongs I'll. Do a quick introduction of that and then also introduce you to the api management product uh three scale.

B

You know how the high level architecture is and how and what the capabilities are, and then I will hand over to the experts, meguel and agueski to take you through the operator capabilities, and you know a demo.

B

We should have some time at the end, about 15 to 20 minutes or so to do a q a session, so we hope to have some good questions at the end of it.

B

Okay, so let me take you through the introduction to the three scale: api management, so, first of all, red hat integration right. So what is integration? So integration is about connecting different applications together, uh whether it is through using uh you know, data through using uh apis through using application uh application uh code right. So there are different ways or that you could connect and uh red hat integration portfolio brings together all of the different ways of connecting these discrete applications. Together in the new world of microservices, we we have a multitude of different.

B

uh You know applications that are going to be running. You know some uh on the cloud, some natively, some uh your older uh applications, and so there is.

B

There is a lot of responsibility for bringing that together and creating uh you know an approach to integrate those different applications together in your solution, so whether you use apache camel for your for developing your core integration patterns or whether you use kafka for for your event, driven and streaming related uh uh messaging activities or whether you use data connectivity and change data capture with the bcm, you would have different options to connect these different integration pieces together.

B

Api management uh now that we are in the api economy- and you know all of the different applications that you interact with expose their their capabilities as apis apis and api management have become one of the core aspects of integration itself.

B

Okay, so here is where we come to red hat's, offering, which is uh three scale: api management and, as with all other products of red hat, it is 100 of open source and we take a modular approach to developing uh this particular platform.

B

So it's uh it ensures that there is no single point of failure and you have the flexibility to uh take the different modules together to and deploy them uh at in different places, whether in the cloud in a container uh uh natively uh you know or- and there are different options uh in which you could run that- and this just ensures uh the different modules uh interact with each other, using the means of apis, which ensures that there is no single point of failure at any point and uh you could at the same time provide a highly scalable solution.

B

Okay, so if I show you a very high level uh high level uh competent uh for three scale, uh you have a control plane. That is the api manager, and this is where you do all of the configuration. So you as an api provider, will have access to an admin portal and an admin api with which you could uh configure your different applications.

B

uh Do your application management to your account management for the different developer accounts that are going to interact with your apis, and you also have a complete cms that would allow you to customize and create a developer portal for your consumers. To actually do some.

B

You know, branded interaction with your apis, have access to sign up to do self management of their keys and the different application plans and billing plans that they sign up for so this is where all of the control level activities are taking place in the api manager, the gateway is your data plane and the gateway is a very thin, uh very thin, uh reverse proxy between your api request and your api back-end.

B

So uh the way of interaction between your gateway and the manager is a very thin uh api listener that is exposed from the back end, which allows you to do authorization of the request.

B

Api requests that are coming in from the various developer, apps that are sending the request, ensure that you know they have the right security to actually access the backend before you allow the request to be processed by the back uh process to the backend and get the response back to your consumers, okay, and at the same time, you also need to report the traffic so that you use it for your analytics, for your rate, limiting and for your billing purposes within your api okay.

B

So uh this is how, as a api api, three scale it with the three skill api manager when you log into the admin portal. This is how it looks on your dashboard. Okay, so you have at a very high level. You could look at your audience. Look at the number of different users and user accounts you have, but at the same time you also have an ability to look at your apis.

B

So here there are two particular terms I want to introduce to you, which are products and back-ends product is the api that you are exposing to your end. Users to your consumers and back-end is the api that you are trying to protect. So back-end is the actual api that is going to be serving the actual response back to the user and your product is uh your the end point from three scale that you are exposing to your consumers?

B

Okay, this is how it looks like and because, as I said, 3skill is an open source project, so the project is available on github and you could see here.

B

If you go to 3scale.github.io, you could see the project it's driven by red hat, but it's open to con contributions and there are a lot of other contributors outside of red hat that are contributing uh to the project now so at the primary level. If you look at the the way that that you know the api manager works now you could see here. uh This is the the same level of complexity that we have shown in the api manager uh and api gateway uh gateway diagram earlier. So the api cast is the api gateway.

B

The branded api gateway from three scale, which is an nginx based api request response traffic. So this is the data plane gateway that controls the traffic between the request and the response, and then on the manager.

B

We have two different modules: the episode module kind of uh is your back-end listener module, so it it exposes the service management api for uh authorizing and reporting the api gateway traffic and then porta is your api manager platform which hosts your developer portal, your account management and your analytics apis, okay, so the api manager, you could actually have multiple deployment options for it, so you could take it and deploy it.

B

You know in multiple modes, so, first of all, it is available uh as part of red hat integration. So you could take three skill and deploy it on openshift, which means wherever openshift could be deployed, whether it's physical virtual on a private cloud, properly cloud or even on openshift dedicated as a managed cloud. You could take three scale and you could deploy the api manager there and manage it there.

B

The other option that you have is to have the three scale hosted option, which is where you take uh here. We have red hat, which is hosting your api manager and all of the back-end components that you require uh in a sas model. So it's just free to sign up so go into threescale.net sign up, get an account, create your apis. Don't worry about scalability and managing the services and upgrading red hat takes care of all that for you, and we are also coming up with a managed option.

B

So for some of you users of openshift dedicated, you know you are familiar with the concept of uh you know off of an sla provided by red hat for an openshift cluster that you get access to by signing up to ost.

B

Similarly, we want to provide you with the managed service option for api management, where you could sign up, it would be available on openshift dedicated you sign up and you get access to in three scale: api manager instance and you would have slas and you would have all of the operational capabilities taken care of by red hat itself.

B

Okay, so the advantage of using three scale on openshift. Is you get a multi-tenant option by default? Which means you get logically separate environments using the same shared resources, so you get a master account with which you could log in as a master admin.

B

You could create multiple tenants and then each tenant would have their own admin portal and they would be able to expose their own developer portal, their own branded developer portal to their users and they'll manage their traffic, their own set of apis and their own set of developer accounts and their own set of api gateways that they could use to manage their traffic in an isolated way without sharing it with the other tenants that are that are going to be using the same environment, okay and similarly for the api gateway.

B

There are also multiple distribution options. So if you remember the gateway, is the data play at the is at the data plane access the policy enforcement point for your api request, so uh api cast, which is uh our api gateway project, allows you to be available as a hosted model as a you know, as a template or an operator based deployment on openshift and uh available to be uh available as a docker formatted container image from the red hat registry itself.

B

So using one of these options, you get the api cache distribution from red hat, which means you get the ability to use the policies, and you know, use all of the functionality that the gateway provides out of the box.

B

You are another important uh policy enforcement point is the three scale uh istio mixer adapter, which allows you to actually uh have uh three scale uh available as a plug-in to your service mesh and allows you to use the same data plane within istio and provide provide that ingress control using the istio ingress router at the same time providing your ability to to manage your service, your policies rate, limiting and billing options within the three scale: api platform itself.

B

Okay, uh so in terms of our our ability and our nativity of working on openshift and on kubernetes platforms, uh we bring in uh the ability to uh in to connect with and expose apis with, k, native and istio. So your your different uh serverless ingress routing or the istio ingress routing ability for you to expose those as apis. How you could do that natively. We provide you with high available and disaster recovery options uh within uh fully fully containerized within the openshift cluster.

B

uh The ability to discover services that are deployed on kubernetes, using just the service annotations, using the api endpoints that are visible within the openshift con console and taking some contextual actions based on the api endpoints that are visible and the use of operators and the operator-based approach for your installation upgrade and maintenance of your api gateway and also, uh in addition to the install of the operators.

B

Also taking care of more capabilities called the application capabilities and using the openshift based approach also allows you to use prometheus and open tracing for your metrics and for your transaction related data and monitoring so which really helps you provide a proper integration with openshift. Okay. So now what I will do is, I will hand over to miguel, to take you through the three scale operator and the application capability.

C

Thank you, satya. I'm going to share my screen now.

C

Okay, hopefully you are seeing the presentation again, okay, so me and nigusky. We work on the operators team in in three scale and at the moment we are offering mainly two operators, three scale operator which allows you to manage three scale: api management environment, and we also offer api cast operator which allows you to manage our api gateway called apicast. uh For this presentation, we are going to focus on three scale operator and we are going to specifically focus on uh application capabilities of the operator.

C

So when I talk about uh application capabilities, what I'm talking about? uh Basically uh application capabilities is a part of the functionality of a three scale operator that allows a user interacting with a three scale. Api management solution, and it does so by exposing familiar kubernetes custom resource objects that allows you to interact at application level with 3scale.

C

But before going deeper into that, I will summarize what functionalities does 3scale operator offer. Mainly. We offer two main functionalities in three scale operator. One is the installation capabilities that part of the operator takes care of, allowing you to install an api management solution uh through a custom resource that we call api manager and it allows you to install that solution.

C

But you also can maintain it over time through this functionality, and we also offer automated upgrades for it for newer versions and changes, and you can manage the lifecycle of that deployment with it, and we also offer several kinds of insights like uh some predefined, graphana dashboards and prometheus alerts and other kind of insights and, on the other hand, there's the application capabilities part of the operator functionality, the threshold operator functionality, which is what we are going to focus in this presentation, and this part of the three scale operator functionality allows you to interact with uh what three scale offers at application level to.

C

It allows you to manage your apis at application level, and for that we offer several custom resources like you see here written. So what features? Does the application capabilities functionality part offers?

C

As I said, it allows you interacting with a three scale: api management solution and with this functionality you can also interact with our our three scale: uh support as a service uh platform. It does not. You do not really need. If you don't want uh to install through scale using the three scale operator, you could just directly connect to sas. So it's a nice additional option that we have, and if you install the three scale operator, you could decide just using the application capabilities. Part.

C

The installation capabilities part is not needed if you already are using sas. uh So, for example, you could install uh your three scale, installation using three scale operator in a cluster and then, for example, use the capabilities part of the operator in another cluster, connecting them both or manage multiple three scale installations with this application capabilities functionality, and so on.

C

Another feature of this uh application capabilities functionality is that it lets. You manage your three scale: application declaratively, so you can have your declarative definitions defined in a given, for example, github repository, and you can integrate it with a continuous integration, continuous development workflow and it uses a familiar interface which are kubernetes. Custom resources objects. So you can manage your api in a kubernetes native way.

C

uh Of course, another nice thing about having a declarative resources is that you can version control them, so you can keep track of the changes over time and you can prepare them and so on as you want.

C

So how does this application capabilities model? Look it's the one you see here, you can think of each of these boxes as a custom resource object. Mainly, you can define your tenants, your logical separations of your api and in each tenant. You can uh define multiple products, which are the your uh user facing apis, that you are defining in in 3scale, and you can connect these products, these these user, facing apis to multiple back-end apis.

C

So for a single product you could be connecting to multiple back-end apis and of course you can do the usual uh work that you do in a three scale: installation, as you can see here, you can define your metrics to keep track of useful statistics and you can also apply limits.

C

uh Pricing rules if you want to do some kind of monetization and you can relate them as you want.

C

Some of the same decisions that we've taken for uh this three scale operator application capabilities functionality is that the configuration source of truth belongs belongs to the kubernetes custom resources.

C

So what you defined there is what it's made effective in your three scale: installation and changes that you might have done on the ui or overwritten by what you have defined it in this custom resources, because the custom resources are the source of truth. uh Another design decision that we've taken for this functionality is that we leverage uh the apis that uh three scale offers to manage your configurations, so the operator is technically capable of doing what three scale apis provide.

C

So if you want to use this application capabilities functionality to interact with your three scale, installation and at application level, what shall I do as a user? So what you have to do is you need to have a running three scale: installation with a tenant account it can be in uh in our three scales as offering or it could be in our on-premise installation installed with free skill operator.

C

Then you have to uh install it. The three skill operator using operator lifecycle manager. We offer a product version and we also offer a community version for it, and then you have to create some credentials to be able to communicate with this three scale: installation through the operator, and then you can go ahead, creating your desired tenants, products and backend using the custom resources that we offer in the operator. So you can manage uh your api at this level, uh but uh so you can see it in a practical level.

C

Now I'm going to handle the to hand over the presentation to gusky and it will show you in action.

D

Thank you very much miguel. So let me present.

D

Okay, can you see my screen.

C

D

Okay, so, as miguel said, the first thing we we need is um a running three-scale installation. This is um some three-scale tenant in our sas platform.

D

So here we have the we have some products and back-ends and what I'm going to show basically is how to create and how to create how to create products and back-ends and how to relate them and how to change several settings of of those. So uh here we have the three scale portal, so usually uh the way of creating products and backends there are. uh There are two ways of of doing this. One is using the ui here. You see, create product or create backends.

D

Here I can create my products and back ends, or I can use the three scale api. It's risky api. It programmatically allows you to create your backends and your product. This is an example. This is the documentation for creating a backend doing a post to this endpoint, providing some information. You create back-ends programmatically, using using your your favorite programming language.

D

What we show here in this demo is a third way of creating not only creating but managing your products and back-ends using uh kubernetes native way of doing that. So you, the the the user uh interface to that is the openshift, the openshift interface, but we need a bridge and that bridge is called the three scale operator. So first thing we need to do is to install the three scale operator. I'm going to install three skill operator. You go to the operators operator app install in three scale.

D

It will take few seconds to do this once the three scale operator is installed. There are new custom resource available for you to manage your three scale products and your three scale backends, it's already installed, as you see now I can.

D

I can uh create my back-ends and I can I can create my products right before creating back-ends and products. I need to create a secret with the credentials to my uh sas accounts to my three scale. Sas account this in this demo. We are showing uh how to manage my tenant in the sas account, but, as miguel said uh previously, it can be your three scale, installation on prem or or any other uh in the same cluster or any other cluster.

D

uh So first thing is creating the the the secret everything is documented. Let me just create the secret, create new secrets. Demo.

D

Here, what I'm doing is is selling the credentials of my three scale tenant to the three scale operator.

D

Now three scale operator knows how to access my three scale tenant. Now, I'm able to create my back-ends and my my back-ends and my products, let's create, let's go to the three-scale back-end, let's create a new back-end.

D

I'm going to delete this, I'm not going into details of all the the attributes of backend. uh I think it's important to you to know that uh I'm creating a backend using the open shift ui.

D

uh So through the operator, we are interacting with our running three scale, so I'm going to create a new backend called demo operated, backend a which is not here, I'm going to show the the.

B

D

Scale dashboard is not here in a in my list of back ends, I'm going to create it so I hit create here we go. I have to wait for a few seconds condition say condition. Synchronized then should be done. Let's refresh my freeskill dashboard, and here we have my operated backend a so uh you can see some of the settings. For instance, the um the the main the main attribute of my backend is the the private base url, which is the my private api url. This is an dot com.

D

If I go to my three scale, dashboard right here I go to my newly created backend and I see that the private base url is correctly set. So let's create another backend.

D

I can create as many buttons as I want. I'm going to create another one.

D

Using the openshift cli well in this case the the ui, but it can be the cli, of course as well. So this is a demo operated back-end v.

D

This is another back-end, so here and waiting for few seconds, the operator takes care of everything to create my uh newly created my to create the backend and show in the in the dashboard of my three scale platform.

D

Okay, here demo operated back in b, so I have two back ends created so now, uh let's create a product from my open shift uh control, so I go to the three scale product. I see these uh custom resource definitions from my uh three scale operator. In this case the fiscal operator is some depth testing operator, but those crds are currently available in the latest risk release which is 2.9.

D

So let's create a product.

D

Jammel so I'm going to create.

D

I'm going to create a product, okay, I'm going to wait for a few seconds. The operator is doing all the synchron synchronization. So it's done this. This state means that um the the the product in three scale is not yet complete. It needs to be promoted for the gateway to to be available, but this is three scale uh internal details. But if I go to my my my uh custom resource, if I go to the conditions, it says that they see the synchronize has been done, so I should see it in my three scale.

D

Dashboard, if I refresh here, we see demo operator product. So if I go to I, this is the the uh three skate portal. uh If I go to my back-ends, this product doesn't have any back-end, so I can add back-ends using openshift as well. I go to my to the jammer definition of the of the product and I can create some. I can create some uh backend.

D

I can connect to previously created backends okay, so it's like I defined some back-end usages, which is the demo operated back in a and demo operated back in b, with some path, some routing path. Okay, so if I hit save I'm again, I'm using up openshift ui, I'm not using um three scale ui or three scale api.

D

So if I go to the product details, it says synchronize, and here, if I refresh in my three scale, operate in my cisco portal, I see that my product has already been connected to the previously connected created backends.

D

So I can show you again. This is my demo operated product created using custom resources. I go to my backends and here you see um that's all I mean I could continue updating, attributes and other uh functionalities, but I guess for the purpose of the demo is showing and once again uh that there is a new way of interacting with your prescale, which is kubernetes native way, using custom resources, and this is available or provided by the three scale operator, application capabilities.

D

So that's all funding over to satya.

B

Yeah, thank you very much. uh That was a really good presentation um for those of you who have not seen three scale before uh you know.

B

A big aspect of actually creating products and back ends was to actually uh either interact with your you know, with your admin portal or trying to do things through the admin ui so really using the operator- and you know using your operator crds as the one source of truth is a very big step and we believe the right approach in trying to bring together uh all of the capabilities of managing and operating three scale.

B

uh You know with your uh you know with your kubernetes native approach, uh karen, I don't know if you want to go through the questions.

A

Yeah and thank you all three of you, because I was having some really interesting side chats on what you were discussing, and you know after this be good to discuss some other things with you. um So let's go back through the questions and tamas uh moss you're. I don't know if you want to hop online or you're. Welcome to turn your camera on too you're answering a lot of the questions in chat, and I want to make sure that um everybody else that's not in the blue jean session can get those too.

A

So, let's start are these apis only for internet access or would it be used for intranet or both inner and intranet I.e? What value does it bring to an intranet based solution?

A

So I know if somebody else wants to join in or tomas feel free.

A

E

B

That sorry yeah, oh sorry, tom, if you're the yeah thomas, if you're there, you could answer it. I didn't see you online. There.

E

B

It doesn't matter, let me.

E

Also enable my camera, um how long yeah, so what I was saying like like uh like we have both like so there's I'd, say a lot of our on-prem customers use the use three scale to manage their internal apis like so, for develop internal developer teams that using each other's apis to to create solutions for their customers.

E

But we also have lots of customers who have public apis where the developers are not part of their company and but in both cases they want to put limits on methods, um and they want to know how much traffic they're getting these kind of things, and maybe in the public cases where the api is public, it's more common to say, okay. I also want like a pricing model, or I want people to sign up on on different kinds of plans.

E

So you have like some some other things you might want to do there, but both prophetic.

A

Nice thanks, um let's see aws api gateway is equivalent to the three scale api. The question is: are: are they equivalent? What's the difference between those two.

E

Yeah we that's not the same. I guess if I understand the question correctly, and so we have frisco has their own gateway, which is called uh ap cast apicast. We always say in spain um um we're in spain, most of us, uh so we say happy customer yeah do we say application api yeah, we say happycast, but most other people new to it would say api pass, and so anyway, it doesn't matter.

E

But uh so that's not the same, but it's so it's so it's um I guess functionality-wise it could be the same, but it's it's it's a gateway and it's a three-scale uh gateway. I'm not sure if anybody with more technical knowledge wants to jump in there.

B

I I think I want to jump in yeah a lot of a lot of people who are new to threeskill, frequently make this comparison with other gateway and gateway providers. What we find is 3skill is a complete api management platform. It's not just the gateway, you know, it's a gateway is one part of it. It is at the data plane, but you also get so much more functionality from the manager itself, like the admin portal, the apis, the developer portal, the self sign on and uh from what I know. Looking at uh the aws gateway.

B

It is a pretty new product and they're, mainly focusing just on the gateway aspect of it gateway and rate, limiting and applying some basic policies, but they don't have the whole ecosystem around building an api marketplace or building an api infrastructure that you could do using three scale.

A

Yeah all right so- and I like that you said abbycast, because that sounds so much happier than api cast. So I love that yeah all right. So how would three scale be deployed for edge use cases? I could see it used at the edge and also in the data center, who wants to jump in.

E

Yeah my original answer there and please correct me: if it's wrong, um it would be to say okay, so what you one thing you could do is have your api cast or api costs uh at the end and set them up in such a way that caching is done. So they don't not every request that hits a happycast has to go all the way to a visinator to actually check if that request can go through if it was rate limited.

E

If the authorization is correct, but you know you can catch this kind of stuff that kind of stuff. So that's what you want to do right, you're, happy cause at the edge and the rest can be like uh in the center.

B

Yeah, I think, when I talked about uh you know the the architecture of it. That's the beauty of it right. Your manager can be hosted in your data center and your gateways could be hosted wherever your services are, so you could go to the edge. You could go within uh your data center or into a different, uh even in a different container platform, or even as a binary, which means that you are not limited to just using uh the gateway that is available with the manager in the same cluster.

B

You are pretty much free to host it anywhere, and all it needs is to make an api call to the back end to validate and report the data, and you could uh you know you could as well do it with caching and with batching policies. You could not even have to do it for every request. You could always uh you know you could always cache these.

B

These authorization data.

A

And you, miguel or xvi say your name, but um do you want to jump in.

A

D

No, I think it's pretty clear from what they said.

C

D

A

C

I wanted to add a little bit more thing. Let me share my screen for a.

C

Second, so uh what we just talked about, uh we have more information, obviously uh about three scale: operators: you can access our repository and we have several user guides uh to quick start you uh into this and some references. So uh you can access here just wanted to add.

A

That's awesome: can you leave that up for a bit yeah, that's great! Thank you all right. um How large is the three scale operator footprint? I know we have a lot of discussions about footprint and openshift, so.

C

So yeah, as I said, uh three scale operator is split into two main functionalities. One is the installation part functionality and the other one is uh the capabilities part. So if you don't need one of these parts, you don't need to install it, so the resources are not used for that part. If you do not use it on the on the operator, footprint is quite small because it's only processing events.

A

That's clear, yeah. Thank you very much all right. We have a comment or a question from youtube when we need, when do we need to use tree scale? What is the sample use case dive into some more sample use cases.

C

Or satya, do you want to? I mean yeah.

E

So um let's say I have this api and I have a large database with recipes and I think there are great recipes and you know I'm gonna make money uh selling those recipes to and other people are going to make apps that actually access those recipes. So I want them to be able to sign up, maybe try on a free trial plan, but then, if they hit 10 recipes, they have to switch to some paid plan.

E

I want to make sure that they don't do a thousand calls per minute against my favorite recipe, because then my server might go down and I want the developers making those apps to be able to sign up get a key to access. The api uh read the documentation, and that would be like the most simple use case. I can think of to use api management. Any api management, like you, want to control, what's going on with your api and you want to make it easy to use it.

C

Yeah, it does not necessarily have to be, for, let's say public users: it could be within your company, so you can control what you. What employees in your company, for example, uh are able to access to a given api that you are developing or some services that your company is offering internally. So you can put some limits and keep track on who's using this api or this service and this kind of rules. So it does it's not necessarily needed only for or used only for public uh users.

C

E

Obviously, not increasingly not, I would say.

A

Feel free to stop sharing yeah like.

C

A

um It was doing this weird thing. um Actually, since we're talking about use cases, I was wondering if each of you could go through and say. Maybe what your favorite use case is. I mean you spin up the cluster and you have to test out what you're working on? What do you usually go to.

B

I could I could start on this first, because uh you know I was actually teaching api management. uh You know in uh you know for three skill: api management for the past three years, so uh our go to use case has always been. uh You know what thomas mentioned right. You have an api right and everyone knows how to implement their api from a given spec right. You and you expose the api. Now. Suddenly you have a public url where anyone can hit the api and access it right. Now. What do you do there?

B

You, you want to put some security on top of it right and yeah. You could go and set up. You know some sort of you know an authentication on the api itself right and then you realize, oh man, I now that this api is available. There's a bunch of people hitting it, and I don't know who, where the traffic is coming from, but my server's down right and then you need to figure out.

B

Okay, maybe what I need to do is I need to set up limits for each user right, but then you say: hey I put limits now, but I don't want everyone to really access my api, my api backend. I want to keep a tab on who can access it. I want them to explicitly sign up and access it based on a specific key.

B

I give them right and once you do that you say hey, you know now I have hundreds of users for this and I you know I want to put different limits for different users. Maybe I have partners, I want to allow them a thousand uh hits per day, uh whereas you know I have some freelance developers.

B

I just want them to hit 100 a day right so now for the same api for the same set of methods, I need to have different plans right and then I I realize hey once they hit this uh this barrier of a thousand hits a day. I want to start charging them, so I need billing plans right, and this is not uh you know a unique scenario. Every company that tries to actually start with uh you know exposing their apis and create an api economy.

B

Out of uh you know the services that they provide kind of hits all of these steps, and so our getting started documentation. I will probably provide this link to uh to in youtube as well.

B

It actually goes through the process of starting with an api and then adding each of these bunch of steps to make sure that you know you're you're, building your api management on top of the apis that you expose- and this is my favorite way to explain- and you know- show the simple use case about what uh how you should use 3scale with your apis.

A

So how granular can you get with that? Is it for you.

B

You could do at a method level you could actually go uh uh have set up an application plan which says you know I would allow get requests, but I will not allow post requests for a particular user or I would allow it like to get requests per minute and zero post request right or you know I could enable and disable at the method level at the path level, uh and I could control uh pricing uh for different, like I could charge for just the post request with a particular plan or just to get requests with the particular plan.

B

So yeah you could get quite granular and your the way you set up your plans is, uh uh you know how you how you make it, depending on the strategy.

E

But it's, but it's just like it's mainly business level uh like the limiting part of it. It's it's, the business loving, limiting more than the operational limiting, which would be like.

A

Yeah, because I was wondering about, can you limit it per name space or have you thought about multi-tenant clusters or how you would do it in multi-tenant scenarios? Things like that.

A

Maybe that's another discussion. Yeah.

B

Yeah, I think yeah yeah.

E

B

E

B

Us an api is an api, whether it is running in one named place or one cluster or a different cluster or even hosted outside of openshift itself right. An api is an api, and so all of the functionality that 3skill provides to do rate limiting and metrics and billing is applicable to all of them.

E

But maybe want to add, like so with the products and the backends. What you can do is like. Let's say you have certain core apis that that you that make sense within multiple api products, then you can have like. So I have one back-end api that I can reuse in multiple products right and I can even uh already save that back end with this.

E

With a certain set of um you know of methods and then for each product, I can define new limits or rules how that packet can use, can be used in that product, not sure if that's partly an instrument.

A

And yeah um all right, we have another question: how does appycast api gateway compared to envoy based api gateways?.

E

But so we're here with the operator team, not with the application you invited.

D

The wrong team.

E

D

A

We'll save that one for the operator team.

B

Although kyle.

A

If you're the kyle, I'm thinking of, uh do you have more yeah.

B

I mean I could answer it really briefly in a minute or two so uh for us right, the the gateway itself.

B

The way that we have uh our gateway uh is uh based on engine x and built on top of open rusty right. So it is a separate component. Now with the onboard based api gateways, a lot of what vendors have started doing is building in the gateway into an envoy itself right and with the with the redesign of service mesh with service mesh 2.0.

B

That is the approach that we are going to take for the rescale istio service mesh uh service, adapter that we that we're going for it is going to be an ny based uh gateway. That is uh also going to interact with the api manager, but then um that that will be specific to the use cases that will be using uh istio or you know, or the similar approach that we will take with courier.

B

But then, though, they don't solve the problem of what do you do for your generic use? Cases where you have a specific http endpoint that you want to? You know manage right. So that's what your api cast does you can do this and you can do that. So it's like a you know, a lot more uh functionality than just an ny-based operator, api gateway.

A

Thanks, thank you. That was a good answer. um Okay, some more questions. Can the operator import a complete free scale, configuration that is api definitions, backend policies, etc?.

D

um I can, I can jump in.

D

Not in a single document, we are working right now, it's not uh it's not implemented. Yet we are working on a functionality which is importing open api spec, although we have to say that the open api spec doesn't cover all the functionalities and configuration that 3scape provides. There are more things related to applications. Application plans metrics. This kind of concepts which which are common in api management, are not covered by open api specification.

D

Although we are working because open api specification is a common known, it's an open specification. So we are glad that customers can work with open documents to start configuring three scale, although they will need to add more things. We are open to um ideas or specs to to make it easier easier for customers, so yeah.

E

Yes, I'm certain, I'm sure, yeah. I think it's uh pretty sure that what we're the way, what we're looking at for future work is is. It is definitely also in the direction of like what you could call api management as code, so that everything that you that you want to your api management can be covered by probably multiple but like yomo files that describe it and you can have a github and you can go go back in time and you know you have it all defined so yeah. That's definitely one of the ways forward.

D

Yeah and and the life cycle of your api, as a customer I mean your api will have upgrades, will have changes, uh so the idea is to provide an automatic, automated way of hand of managing this life cycle. So you change your api definition, you upgrade it or you add some endpoint or whatever and automatically.

D

uh There is a pipeline that automatically those changes are are configured in three scale, so you don't need to do anything. This is the one of our objectives.

D

We have in three scale one tool, one cli, which is called three scale toolbox that has been used for these purposes, trying to automate the life cycle of your api, and now this functionality we are trying to make it where net is native using the operator. So we are moving some functionalities to the operator.

D

For this regard,.

A

Nice, thank you and we are just about out of time. So that was a really good discussion and thank you so much for joining. Do you have any last things that you want to say in our three minutes remaining.

B

I think uh you know we had some really good questions and from the questions, what I could see is a lot of the audience today is fairly new to three scale, uh so I could think of two ways that you could immediately start exploring three skills, so you could go to signup.3skill.net and sign up for an account and you could play with the user interface and create sample apis and services play with it and get to know the product a bit better and for those customers and those users who are already familiar with openshift or have open shift installations, and I think, getting the three scale operator and getting started with installing it.

B

uh On your. uh You know on premise: installation of openshift, and uh you know exploring. That is probably a good way to do. There are a lot of good documents out there on getting started, and you know, and some use cases that we could point them to, and you know- and I think this is a really good uh product that brings a lot of additional functionality on top of openshift, and you know this will be a welcome addition to your customers.

A

Awesome, thank you and that was signup.3scale.net right, yep, um okay! Well, thank you again and thank you everybody for joining another openshift commons, this time with the three scale, api management team, awesome that was really great and gave me a lot to think about, and I know others I was talking to you offline. So thank you again and for next time.

A

In two weeks we have um index image builder service, um brian cook's gonna, be here from the open shift management, our pm team, and that one will be really good as well. If you work on operators and are interested in catalogs, thank you again and chris. If you.

D

D