xdr-terraform-modules/thirdparty/terraform-aws-github-runner/docs/test-lambda-local.md

Lambda - Test locally

This README provides guidance for testing the lambda locally / and or in AWS. This guide assumes you are familiar with AWS, lambda and Node. If not mentioned explicitly, comments provided should be executed from the root of the lambda package.

Testing in AWS

Just navigate to the Lambda in the AWS Console and trigger a test event. Provide an event that matches the required input. For lambdas that does not require a specific event, just send any event.

Testing locally

Testing locally can be done in two ways; using AWS SAM framework or run via a wrapper to simulate the event to invoke the lambda. Both setups require that the mandatory input environment variables be set, and AWS resources on which the lambda depends are available. We advise for testing the lambda locally to first create your own deployment of the whole module to AWS, this will simplify the setup of dependent AWS resources. For example, based on the de default example.

Local test setup instructions are available for the following lambda's:

• runner-binary-syncer - This lambda does not need any input, no event is required. Supported via SAM and local Node.

Extend deployment configuration

Add the code below to your Terraform deployment to allow your principal to use the Lambda role and retrieve the lambda configuration. Update your Terraform deployment and apply the changes.

data "aws_caller_identity" "current" {}

module "runners" {
  
  ...

  # Assume you have a profile with Admin privileges, allow you to switch to the Lambda role
  lambda_principals = [{
    type        = "AWS"
    identifiers = ["arn:aws:iam::${data.aws_caller_identity.current.account_id}:root"]
  }]

}

output "development" {
  value = {
    lambda_syncer = module.runners.binaries_syncer.lambda
  }
}```

Once you have updated your Terraform deployment you need to read the lambda configuration into your environment. Run the commands below in your Terraform workspace folder.

bash LAMBDA_ENV=$(terraform output -json development | jq -r '.lambda_syncer.environment[].variables' | jq -r "to_entries|map(\"(.key)=(.value|tostring)\")|.[]") for x in $LAMBDA_ENV ; do echo setting $x; export $x; done

### Testing with SAM

This setup requires AWS SAM CLI and Docker is installed locally. First update the AWS config (`~/.aws/config`) so you can use easily switch to the role used by the lambda.

properties [profile gh-development] source_profile= region= role_arn=

Now you can set the profile and region as environment variables or pass as argument to SAM.

export AWS_REGION= export AWS_PROFILE=gh-development

For SAM a `template.yml` defines the lambda for running locally. Thats all, now build your lambda with `yarn run dist` and then invoke the lambda with `sam local invoke`.


### With Node

Instead of using SAM you can use Node with `ts-node-dev` to test the code locally. The drawback is that you have to setup AWS credentials in your shell. Also, you are dependent on a tiny wrapper (`local.ts`), and your local Node version.

The AWS SDK does not seem to handle environment variables for profiles, the only option to pass the role is via credentials. You can get credentials via STS for the role.

bash

role=$(aws sts assume-role

--role-arn "<ROLE>" \
--duration-seconds 3600 --role-session-name "dev")

export AWS_ACCESS_KEY_ID=$(echo $role | jq -r .Credentials.AccessKeyId) export AWS_SECRET_ACCESS_KEY=$(echo $role | jq -r .Credentials.SecretAccessKey) export AWS_SESSION_TOKEN=$(echo $role | jq -r .Credentials.SessionToken) ```

Next set the region export AWS_REGION=<region>. Now you can run the lambda locally via yarn run watch.