Terraform lambda docker

This is an opinionated terraform module that enforces a few conventions in the building and naming of docker images used in AWS Lambda functions.

This module does not declare Lambda functions. It's best used in conjunction with the lambda module.

These conventions are:

1. The dev workspace creates the ECR repository, other workspaces use it (this distinction boils down to resource vs data-source usage).
2. The terraform workspace will be the (mutable) tag of the image.
3. Images are identified by their SHA256 signature.
4. A recomended name for the lambda function will include the workspace. This implies that the mapping between environment, workspace, and lambda is one-to-one. This also implies that using lambda aliases is not encouraged.
5. Terraform code is expected to live in a folder one-level deeper than the lambda code.

Assuming all of the conventions.

module "lambda_image" {
  source = "github.com/ItrendCL/terraform-lambda-docker"

  repository = "github-user-or-organization/repository-name"
}

Using a multi-lambda repository. This assumes a folder named a and a folder named b next to a folder containing the terraform code, as per the last convention above.

module "lambda_a" {
  source = "github.com/ItrendCL/terraform-lambda-docker"

  repository = "github-user-or-organization/repository-name"
  suffix     = "a"
}

module "lambda_a" {
  source = "github.com/ItrendCL/terraform-lambda-docker"

  repository = "github-user-or-organization/repository-name"
  suffix     = "b"
}

Using all of the parameters, and not assuming the last convention.

module "lambda_a" {
  source = "github.com/ItrendCL/terraform-lambda-docker"

  repository = "github-user-or-organization/repository-name"
  suffix     = "some-suffix"
  path       = "." // use this folder as docker context

  image_expiration_days = null // untagged images won't expire from the ECR
  image_build_args      = { VERSION = "1.2.3" }
  image_dockerfile      = "production.Dockerfile"
}