Kotless configuration

Kotless configuration of a Gradle plugin consists of a few globally applied parts:

• Terraform configuration;
• Kotless service configuration;
• Optimization configuration.

Here is a simple snippet of the whole configuration:

kotless {
    config {
        bucket = "kotless.s3.example.com"
        prefix = "dev"
    
        dsl {
            type = DSLType.Kotless
            //or for Ktor
            //type = DSLType.Ktor
            //or for SpringBoot
            //type = DSLType.SpringBoot
        }
    
        terraform {
            profile = "example-profile"
            region = "us-east-1"
        }
        
        optimization {
            mergeLambda = MergeLambda.All
        }
    }
    //<...> - Web Application config and Extensions Config
}

Let us take a look at the items of this configuration one by one.

Terraform configuration

In a terraform configuration you can set up a version of Terraform used, its AWS provider, and a bucket for tfstate.

You will need to set up region and profile — region for deployment and local profile used for it accordingly.

Here is the simplest possible snippet with a setup of Terraform configuration:

terraform {
    //Will use for terraform state buckets used by kotless service configuration
    //version of terraform and aws provider will be default
    profile = "example-profile"
    region = "us-east-1"
}

Note that the chosen version of Terraform will be downloaded automatically.

DSL configuration

In a dsl configuration you may define type of DSL used: DSLType.Kotless or DSLType.Ktor, otherwise type will be deduced from used DSL dependency.

In case of Kotless DSL and Spring Boot, you may change staticsRoot variable — it is a folder that Kotless will use as a root to resolve all static route files. In Ktor DSL it is defined via staticRootFolder variable in code.

dsl {
    type = DSLType.Kotless
    //or for Ktor
    //type = DSLType.Ktor
    //or for Spring Boot
    //type = DSLType.SpringBoot
}

Kotless service configuration

Kotless service configuration is a set of directories, S3 bucket, and several values used by Kotless to deploy your application.

You will need to set up bucket — it is a name of a bucket that Kotless will use to store all files. For example, it will store packed jars and static files there.

You can set prefix variable — it is a prefix with which all resources created in a cloud will be prepended. prefix can be used to deploy several environments of one application.

Here is a simple snippet with a setup of service configuration:

config {
    bucket = "kotless.s3.example.com"
    prefix = "dev"
    //<...>
}

Note that bucket value will be used for Terraform state, if it is not set explicitly in the Terraform configuration.

Optimization configuration

We are doing our best to make Kotless-based lambdas as fast as possible.

There are plenty of optimizations that are embedded in a Kotless synthesizer and runtime. Some of them can be configured to align with you needs.

Lambda Merge Optimization

This optimization defines if different lambdas should be merged into one and when.

Basically lambda serving several endpoints is more likely to be warm.

There are 3 levels of merge optimization:

• None — lambdas will never be merged;
• PerPermissions — lambdas will be merged, if they have equal permissions;
• All — all lambdas in context are merged into one.

Here is a simple snippet of optimization configuration:

optimization {
    mergeLambda = MergeLambda.All
}

Lambda AutoWarming Optimization

This optimization sets up a timer that will autowarm lambda, by default, each 5 minutes. Such an optimization makes cold start less frequent.

Each timer event executes a warming sequence. This sequence triggers LambdaWarming objects and is described in a Lifecycle Control section.

optimization {
    //default config
    autowarm = Autowarm(enable = true, minutes = 5)
}

Kotless's lambdas can be autowarmed. It means that some scheduler will periodically (by default, each 5 minutes) call lambda to make sure it will not be displaced from the hot pool of cloud provider.