AsyncAPI Demos of products using AsyncAPI, 26 Feb 2019

Previous Meeting

⏯

youtube image

►

From YouTube: Kyma Overview of Key Features - Demo

Description

Video with explanation of:
- Why you should use Kyma?
- Demo of key features basing on specific scenario
- Demo of how Azure services are consumed in Service Catalog through the Azure Broker
- Demo of how thanks to Grafana through Prometheus and Istio you get easy access to statistics of your Serverless usage
- Demo of how you can easily consume services in Lambda code thanks to our easy service binding functionality

You can do all of it already with Kyma.

A

Have you ever been thinking? How do you extend your applications in the cloud native world? The answer is pretty simple: its schema. You want to provide extensibility to your application, but stay independent from its release cycle and we have component for that. It's called application connector. You also want to get on speed with delivering new features, just focus on business logic and don't care about setting up operations and that's handled by our surveillance component.

A

You also want to make it easy for your developers to consume functionality, exposed by your application and use best managed services from various cloud providers. That's where our service management comes into play, and last but not least, you don't want to waste time on combining all those different cloud native technologies and have them ready for you. That's where all the rest of components come into play is like promise use for monitoring like Jager for tracing or Easter for service mash, but that's all theory right and nothing can be better than real demo.

A

That will show you all these features in one scenario: I'm gonna quickly show you how you can extend your application in this case. It's gonna be a commerce application where you have a typical functionality of giving comments to a product, and you want to extend the functionality of comment. You know that stars are not enough.

A

If customer gives you one star or five stars, you don't really get the sentiments of your customers and you want to enhance the functionality of monitoring the comments with some good machine learning, so this application will send an event that contains the comment into the kheema cluster and first component. It's gonna touch is the event service from the application. Connector. Then, thanks to the even bus functionality, it will trigger our lambda. That is responsible for comments evaluation.

A

It will hit the asian text analytics service to get the sentiments for the comment and then, if the comment is negative, it's gonna be saved and managed MongoDB offering by a juror. You will also see another lambda. It is listing the negative comments. I'm gonna show you how easy I can access this list of negative comments from outside the key McCluster we're gonna.

A

Do it by triggering the lambda with the HTTP call and the lambda will call the age of service to give me the list of negative comments and at the end, you can see how, thanks to this verbs mesh, we get all the real-time data for the monitoring and how easy it is to access them in this demo.

A

You will not learn how it was possible to establish a trusted connection between external application and our application connector, because that's just a one-time configuration you will also not see the provisioning of the aerial service broker and enabling listing of all the aedra services in the Service Catalog. Now we are inside the kheema console UI. So once you have your applications connected and Brokers provisioned, the first place your developer should go to. Is the service catalog, because in the service catalog, you can find all the services needed to provide extensibility.

A

So, in case of our scenario, first thing I need is to get access to the events exposed by my connected application. So let me filter of the view. I can see that I have two connected applications with services.

A

Our scenario covers Commerce Mach, so let's filter that out and let's see one of the classes I see the service that enables the API it provides the open, API spec, so the swagger is also available, but that's not what I need what I need is a service that gives me access to the events and because the service during the registration provided the icing API specification for the events, I can also see them rendered here in the UI. So that's the first service that I need to provision in my namespace.

A

Okay, so now I need a MongoDB, maybe Google Cloud has some good services. Unfortunately, there's no database for my use case. So let's find something different, let's just clear over the filters and just manually search mumble. Okay, so Ezra exposes the Cosmo DB, which supports the MongoDB API perfect for my use case. There's a basic documentation provided by us and also I have link to additional documentation provided by a juror. This is a perfect service for my use case. Let me add it to my namespace.

A

Thanks to the open service broker, API specification, that is part of the Service Catalog and the brokers I, can see that Asia also provided a specific schema of all the details that I need to provide for a database to get it properly. Provisioned and the last service that I need is a kind of text analytics. Let me try to filter by tag. Maybe a juror is tagging those classes in a specific way, but there are many different tags in here. So let me just narrow cognitive, okay, perfect. So a sure text analysis is available.

A

Everything I need is in the service. Let me add it to my namespace in the lambda UI. You can see that I have two different lambdas comments sentiments counter. This lambda is responsible for checking the comment. If it's negative or not, and then in case of negative comment, it saves it into the the most important part here to notice is the service binding area? So you can see that this lambda is binded to two different service instances.

A

The aja text analytics and the MongoDB I can see the information about the environment variables that, thanks to this binding, are automatically injected into my lambda. Important to notice is also the even trigger so I could easily set it up after accessing the list of available even triggers. The other lambda is pretty similar. The biggest difference is the type of the trigger here. The trigger is the HTTP, but the most important is to see those lambdas in action. Let us switch now to the comment line using the mock of my application.

A

I was able to send three different comments to Kemah here. In the comment line, you can see the information about the logs produced by the lambda, so you can see that the first event that came in was marked as negative. The comment was pretty clear. The second one also was negative, and the third comment was marked as not really negative. Now let us switch back to the UI and, as I mentioned before, we provide out-of-the-box integration of the monitoring.

A

So we can see that the comments sentiments counter has a link to the dashboard, so I can access directly to the graph on a dashboard and I can see all those three requests that came into the lambda and all the most important statistics for the performance of the lambda. Now let us move to the last step of the scenario, the other lambda that is about to list the negative comments it's configured to be triggered by HTTP call.

A

This means this lambda is exposed with help of SEO and I can easily access the data using the REST API. Let me now try to access the publicly available lambda, which is not secured for this scenario, and because we've sent three comments and from the logs we've learned that only two are marked as negative. We should also get a reply with only two negative comments. As expected, we got only two comments in reply. Thanks for watching I'm Akash greens can I invite you for more details or team. A project on your page Cheers.