-
Notifications
You must be signed in to change notification settings - Fork 58
Getting Started
In this short tutorial we will overview the steps of creating a Kotless-based application.
To use Kotless in your project you need:
- Route53 DNS zone to create DNS name for the application
- S3 bucket to store kotless-related artifacts
- ACM certificate for DNS name at US-EAST-1
First of all, you need to set up Gradle for your project. Kotless uses Gradle tasks to prepare deployment environment, generate deployment code, prepare a jar and deploy it to cloud provider.
You need to apply a plugin to project at build.gradle.kts
:
id("io.kotless") version "0.1.1" apply true
To use Kotless DSL in a project you need to set up a maven repository with DSL and add the necessary dependency:
repositories {
jcenter()
}
dependencies {
implementation("io.kotless", "lang", "0.1.1")
}
Next, you need to set up Kotless. Here is a simple configuration with comments:
kotless {
config {
//bucket that kotless will use to store its artifacts
bucket = "kotless-example-bucket"
//prefix that will be added to all resources created in AWS
prefix = "dev"
//directory against which @StaticGet annotated files are resolved
workDirectory = file("src/main/static")
terraform {
profile = "example-profile"
region = "us-east-1"
}
}
//webapp for this gradle project
webapp {
//configuration of lambda created
lambda {
memoryMb = 1024
timeoutSec = 120
}
//packages in which kotless should search for code
packages = setOf("org.example.kotless")
//route53 alias for deployed application
route53 = Route53("kotless", "example.org")
}
}
Let's write your first dynamic route.
In Kotless terminology, "dynamic route" is an HTTP route processed by a lambda. On the contrary, "static route" is an HTTP route mapped to an S3 object.
Creating dynamic route is relatively simple. This code snippet will
create a route at HTTP path /
, which will print "Hello world!"
@Get("/")
fun root(): String {
return "Hello world!"
}
Note that dynamic route function may have parameters. Parameters will be taken from the URL and body (if presented) and deserialized into the parameters of function. Moreover, the result of function will also be serialized to HTTP response automatically.
In case you need a full control of the HTTP response you may return HttpResponse
object,
which will be passed to ApiGateway without any changes.
Most of the real-world applications contain static resources, e.g. CSS and JS files. Serving static resources with lambdas is a very wasteful spending of resources, so Kotless proposes a solution — static resources mapped directly to S3.
This code snippet will create route at HTTP path /file.css
with file example.css
@StaticGet("/file.css", MimeType.CSS)
val exampleCss = File("example.css")
Note that the file will be resolved against workingDir
set in Gradle configuration.
Finally, you need to deploy your Kotless-based application to AWS. Check that
fulfilled all the preliminary conditions and execute deploy
task in Gradle.
Kotless will download Terraform, generate deployment code, pack lambda, and, at the end, deploy the application to AWS.
As soon as the deployment is completed, your application will start serving requests at the Route53 DNS name you have assigned to it.