On today’s episode of Adventures of Angular we’ve got panelists Ward Bell, Joe Eames, Alicia Michael, John Papa, Charles Max Wood, and our special guest Shayne Boyer. Shayne is a Senior Developer Advocate at Microsoft and on the Azure team. The last time he was on the show was Episode 082 of Adventures in Angular and we talked about getting started with Angular 2. Today we are going to talk a bit about Angular and Azure. Stay tuned.

[3:12] Talk about offerings with Azure.


There is a connotation that if you want to use Azure you have to use .NET
That is not the case.
Large part of Azure runs linux.
There are over 170 type products that Azure offers.
noSQL databases, postgres, mongolDB, Azure Cosmos, Azure functions
It’s more than .NET and VMs.
Things you can do this type application are things like - Deployment to web front end, putting apps in Docker container and pushing to container service, scaling those apps etc etc.


[5:23] Put your app in Docker container?


Talking about just front end. Just the web application.
Putting it into a container and deploying the container into a linux instance or web app on Azure.


[6:13] Why is it a good thing to use Docker for this kind of thing?


Example. NGX for front end services, you can’t NGbuild using ClI do an NGBuild get the disc folder and throw that to a web application service like IAS or Node and have that application just service. Deep linking wouldn’t work.
Instead you would want to package something like an express server that requires NodeJS. Then something to handle deep linking. You can easily package that in a container and push it to the cloud and be able to control it without worrying about infrastructure.
Essentially it’s the app that has been written as well as the server that serves that app.
You can choose the base it runs on. I.e. Node 6 instead of 8. Etc etc.
Control those in the container so any time you pull it down it has those same settings.
Often developers assume versions of services are the same between the developer and the services used like a cloud service and when it’s not, things break.
The concept of “works on my machine” is actually true now. How it works for you, will be exactly how it works where ever you push it.
You can set up the version of Node you want, the settings how you’d like, test it on your own machine and when you push it to providers like Azure, Heroku, AWS, etc, it will run the same on all those providers.
Takes away complexities when testing.


[9:39] Outside of Docker, what other things does cloud bring to the table?


Serverless offerings.
Takes away issues with - worry about building a node server to serve my app. Building API to serve the data. Building Infrastructure behind the server and deploying it.
Building and deploying pains reduced as well.


[10:57] “Wait wait, serverless?”


It’s the new buzzword.
There are servers underneath.
Don’t have to worry about infrastructures or the servers themselves.
Just write the function.
Function will return the data to controller service built in Angular. Just have to build the Javascript (or python, and C#, java is coming) in the portal or in Github, and it’s just the code to run the function.
No need to worry about the types of servers, VMs, operating systems, patching or scaling.
It will scale based on what the capacity demand it needs.
Event driven - event queues, message queues, etc.


[13:20] Simple endpoints


Scalability at endpoint level.
Previously when writing APIs on the backend, typically you’re concerned with scaling that API application.
Endpoints typically scale at the same level.
Serverless functions scale at that typical API level.
Paying only for usage.


[15:30] How do you orchestrate between the services?


Just because you spread things around doesn’t mean it’s better.
It’s important to realize that breaking assembly or dll files down into smaller dll files don’t change things much.
Minimizing what it takes to think about when it comes to handle and configur