Kong Kong Konnect Tutorials, 6 Jan 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Rapid Application Modernization With Kong Konnect and Apache Camel

Description

In this video, Simon Green, solutions engineer at Kong, shows us how we can modernize a legacy application in the most efficient way possible to improve the performance and give it a new lease on life in a multi-cloud, platform-agnostic environment like Kubernetes.

Learn more about Kong Konnect: https://bit.ly/3vDuUBz

Links and more detailed info in the blog post: https://konghq.com/blog/rapid-application-modernization/

▬▬▬▬▬▬ TIMECODES ▬▬▬▬▬▬

0:00 Intro
0:17 Code First Approach
1:50 Demo
6:18 Conclusion

#LegacyAPI #APIgateway #ApacheCamel #KongKonnect

A

In this video we're going to review how to rapidly modernize legacy applications with con connect and apache camel leveraging the power of kong's api gateway together with the powerful, camel rest domain. Specific language here is the flow we use for a code first approach.

A

Our developer begins by refactoring the existing application, in this case a python client which sends asynchronous requests to ansible by replacing ansible with apache, camel we're able to modernize our application into a synchronous, restful app that can be containerized and easily deployed to kubernetes.

A

In addition, our developer has woven open api documentation into the camel rest dsl, allowing us to easily import our specification into insomnia from here we're able to initiate an api ops approach using enzo, dec and cube ctl to deploy and expose our modernized application via the congateway.

A

This can easily be integrated into our cicd pipeline and, together with terraform, give us an immutable deployment which can be deployed literally anywhere. In addition, we're able to support existing legacy python clients by using versioning in kong gateway.

A

This method avoids forcing clients to upgrade to the modernized application until they're ready to do so and with that, let's get started with the demonstration.

A

If we go over to vs code, my developer ide you'll notice that here in the camel rest dsl, we have defined a restful endpoint and within that endpoint we have the open api documentation nicely woven in.

A

We also have an http put endpoint that acts as the gateway or the input to for request going to our pig latin translator.

A

We also have a camel route that takes care of some basic transformation tasks with the request, and then we use a java bean to actually do the pig latin trend translation.

A

If we want to test out our camel route in spring boot. We can do this via the command line.

A

Via the cli, we now have a local host endpoint listening on port 8080, and by going over to insomnia our api designer of choice, we can easily import the swagger documentation from camel by hitting that url.

A

Now that we have the open api document imported into insomnia, we're able to validate it using what we call linting. So just by typing any kind of error, we get some kind of error handling through the validation logic.

A

We also have the http put request available to us so that we can send requests to our spring boot endpoint.

A

If we copy over a sample request payload to insomnia by hitting send, we will get back a response from spring boot with the transformed pig latin sentence.

A

In amazon we already have kubernetes stood up using eks, the amazon flavor of kubernetes.

A

For the gateway we are using, kong connect, kong's managed service for the gateway, and we have multiple versions of our service already stood up. We have a version with the original funnels requests over to ansible and then version 2, which funnels requests over to the spring boot camel version which has already been deployed for us, and here is the original python script and the new python script, which represents our clients so first off.

A

If we go over to the scripts directory, we're able to use the original python script to hit our kong gateway, which funnels that request over to ansible and what you'll notice is that it's taking almost three seconds or more than three seconds to get back a response due to the asynchronous nature of the ansible api.

A

Now, if we try the newer client, which hits the version 2 or the spring boot version of the synchronous, restful, endpoint you'll notice that we get back a response within 70 milliseconds, so vastly vast improvement on performance going back to connect.

A

What we have here is the service hub and if we drill down to the service that we are calling, you can see the number of requests that have flown through our endpoint, the number of versions that we've defined, and here in the throughput metrics. You can see that we've had two requests coming from the python client for version two and the rest for the original client, so we're able to basically handle requests both from legacy clients and the new clients.

A

To recap what you've seen today is how a developer's code first approach to application modernization can work in the real world by modernizing a legacy python app with apache camel. We were able to generate an open api specification which ties into our api ops process by using insomnia to mock test. We deployed our application using a combination of enzo dec, cube, ctl and terraform to the kubernetes environment of our choice.

A

Lastly, by using kong gateway to provide api versioning support, we ensured that clients were not forced to upgrade to the latest version until they're ready to do so.

A

Last but not least, by using kong's shared policy layer via the kong gateway, we ensured that developers were not concerned with common policy and governance matters, including routing load, balancing security rate limiting and cashing by centralizing. These concerns with the kong gateway.

A

We allowed our developer to use the technology and language of their choice to see what else you can do with kong gateway head over to konghq.com.