Not able to create AWS Lambda using CloudFormation without code - amazon-web-services

Question

I am trying to set up a resource pipeline, where I want to deploy all my resources using CloudFormation. I have a separate pipeline to deploy code.
Using below CloudFormation template
{
"AWSTemplateFormatVersion": "2010-09-09",
"Resources": {
"S3Bucket": {
"Type": "AWS::S3::Bucket",
"DeletionPolicy": "Retain",
"Properties": {
"BucketName": "devanimalhubstorage"
}
},
"HelloLambdaRole": {
"Type": "AWS::IAM::Role",
"Properties": {
"RoleName": "HelloLambdaRole",
"AssumeRolePolicyDocument": {
"Statement": [
{
"Effect": "Allow",
"Principal": {
"Service": "lambda.amazonaws.com"
},
"Action": "sts:AssumeRole"
}
]
}
}
},
"AnimalhubLambdaFunction": {
"Type": "AWS::Lambda::Function",
"Properties": {
"FunctionName": "AnimalhubLambdaFunction",
"Role": {
"Fn::GetAtt": ["HelloLambdaRole","Arn"]
},
"Code": {},
"Runtime": "dotnetcore2.1",
"Handler": "myfirstlambda::myfirstlambda.Function::FunctionHandler"
}
}
}
}
Problem Resource handler returned message: "Please provide a source for function code. (Service: Lambda, Status Code: 400, Request ID: 377fff66-a06d-495f-823e-34aec70f0e22, Extended Request ID: null)" (RequestToken: 9c9beee7-2c71-4d5d-e4a8-69065e12c5fa, HandlerErrorCode: InvalidRequest)
I want to build a separate pipeline for code build and deployment. Can't we deploy Lambda function without code?
What is the recommended solution in AWS? (I used to follow this approach in Azure, am new in AWS)

Answer 1

First, I would advise you to look into AWS SAM. It is very helpful when creating and deploying serverless applications and will have a lot of examples to help you with your use case.
Second, using separate pipelines for this purpose is not the recommended way in AWS. Using dummy code, as the other answer suggests, is also quite dangerous, since an update to your cloudformation would override any other code that you have deployed to the lambda function using your other pipeline.
In a serverless application like this, you could make a separation into two or more cloudformation stacks. For example, you could create your S3 buckets and other more "stable" infrastructure in one stack, and deploy this either manually or in a pipeline. And deploy your code in a separate pipeline using another cloudformation stack. Any values (ARNs etc) that would be needed from the more stable resources, you could inject as a parameter in the template, or use the ImportValue function of CloudFormation. I'd personally recommend using the parameter since it is more flexible for future changes.