Lesson 1: Customizing a Web Server

In lesson 0 we built a simple web server. In this lesson, we will build on top of this web server. We will add the ability for the user to run the web server in the command-line and to configure it through flags and arguments.

This kind of component is something you will see in multiple projects in the Kubernetes community. One great example is the Kubernetes API server.

So let's get to it.

Knowledge bits

• we can reference packages that belongs to the standard library like "os" directly by adding a line in the import section.

  And we can also reference external packages like github.com/contributing-to-kubernetes/go-examples/lesson-001-web-server-cli/app directly

• in this lesson we will be using a variant of the nil pattern explained in lesson-00:

	command := app.NewServerCommand()
	if err := command.Execute(); err != nil {
		log.Println(err)
		os.Exit(1)
	}

Here we use the definition of err and the condition to change our program workflow in the same line.

• after checking the actual code from this lesson, you might be wondering "where are those -h and -v flags come from?"

  Don't panic, those are flags added by default by the cobra library as explained in the overview section of the documentation and and in the version flag section

Getting Started

In order to build our web server as a CLI we will use https://github.com/spf13/cobra.

The main addition to our web server will be the definition of a cobra command to define flags and receive arguments.

Testing It

Since we have a CLI, we now have useful descriptions and help messages

$ go build -o server && ./server --help
The server is an example application to mimick the organization of the
Kubernetes API server.

Usage:
  server [flags]

Flags:
  -a, --addr string   server's address (default "0.0.0.0:8080")
  -h, --help          help for server
  -v, --version       version for server

You can see we have some flags. Let's see what version our app is

$ go build -o server && ./server -v
server version v1.0.0

And to run it

$ go build -o server && ./server
2020/05/17 14:42:57 version: v1.0.0
2020/05/17 14:42:57 args: []string{}
2020/05/17 14:42:57 starting server at 0.0.0.0:8080

And finally, if you want to specify a different listening address for the server

$ go build -o server && ./server -a 0.0.0.0:9090
2020/05/17 14:43:40 version: v1.0.0
2020/05/17 14:43:40 args: []string{}
2020/05/17 14:43:40 starting server at 0.0.0.0:9090

You should be able to test it by going over to http://localhost:9090/

Next steps (optional)

• we can use viper library to make use of another common pattern which is reading configuration from files.

For example having a file with the server address:

hostname: "0.0.0.0:8123"

• accept more flags to the CLI. Use and see how to deal with required flags, optional ones.

• we can think about how to make our program more resilient and for example assume defaults if a required flag is missing,..