aws-cdk: success and failure lambdas for state machines are being duplicated in iam roles

[theomantz-luna]

Describe the bug

success and failure lambdas for state machines are being duplicated in iam generated roles.

Expected Behavior

the success / failure path lambda should be added to the associated policy once.

Current Behavior

when creating a success and error lambda for each lambda in a state machine, each success and failure lambda is added to the iam policy document. even when they are the same lambda. due to this, iam policy document size limits are reached quickly as each document that has success and failure lambdas associated with it has n+(n * 2) resources added (where n is the number of unique lambda functions).

Reproduction Steps

create a definition body with the following general format:

DefinitionBody.fromChainable(
            sfnChain
                .addCatch(
                    errorReporterTask.next(
                        new sfn.Fail(this, name + " Fail State"),
                    ),
                )
                .next(successReporterTask),
        )

Possible Solution

perform a minimal compression on the iam policy document before deploying so that lambdas with duplicate names are not added to the same policy document

Additional Information/Context

No response

CDK CLI Version

2.133

Framework Version

No response

Node.js Version

v18.20.2

OS

macOS (silicon)

Language

TypeScript

Language Version

5.4.4

Other information

No response

[pahud]

Hi

Can you elaborate more about the issue?

This is my testing code

export class DummyStack extends Stack {
  constructor(scope: Construct, id: string, props: StackProps) {
    super(scope, id, props);

    // dummy lambda function
    const fn = new lambda.Function(this, 'CatchFn', {
      runtime: lambda.Runtime.NODEJS_LATEST,
      code: lambda.Code.fromInline('foo'),
      handler: 'dummy.handler',
    });
    const fntask = new sfntasks.LambdaInvoke(this, 'CatchFnInvoke', {
      lambdaFunction: fn,
    });

    const fn2 = new lambda.Function(this, 'NextFn', {
      runtime: lambda.Runtime.NODEJS_LATEST,
      code: lambda.Code.fromInline('bar'),
      handler: 'dummy.handler',
    });
    const fn2task = new sfntasks.LambdaInvoke(this, 'NextFnInvoke', {
      lambdaFunction: fn2,
    });

    const start = new sfntasks.LambdaInvoke(this, 'StartState', {
      lambdaFunction: new lambda.Function(this, 'StartLambda', {
        runtime: lambda.Runtime.NODEJS_LATEST,
        code: lambda.Code.fromInline('start'),
        handler: 'dummy.handler',
      }),
    })


    const definitionBody = sfn.DefinitionBody.fromChainable(
      start
          .addCatch(
              fntask.next(
                  new sfn.Fail(this, " Fail State"),
              ),
          )
          .next(fn2task),
    )

    new sfn.StateMachine(this, 'StateMachine', {
      definitionBody,
    })

  }
}

And I got this in the default policy of the state machine execution role:

{
	"Version": "2012-10-17",
	"Statement": [
		{
			"Action": "lambda:InvokeFunction",
			"Resource": [
				"arn:aws:lambda:us-east-1:<DEDUCTED>:function:dummy-stack2-CatchFn9B5768B8-ga1bHwyKMnlm",
				"arn:aws:lambda:us-east-1:<DEDUCTED>:function:dummy-stack2-NextFn2EC1E31E-2qr1IoR8eTSX",
				"arn:aws:lambda:us-east-1:<DEDUCTED>:function:dummy-stack2-StartLambda81613300-GgkyN4bUAdvo",
				"arn:aws:lambda:us-east-1:<DEDUCTED>:function:dummy-stack2-CatchFn9B5768B8-ga1bHwyKMnlm:*",
				"arn:aws:lambda:us-east-1:<DEDUCTED>:function:dummy-stack2-NextFn2EC1E31E-2qr1IoR8eTSX:*",
				"arn:aws:lambda:us-east-1:<DEDUCTED>:function:dummy-stack2-StartLambda81613300-GgkyN4bUAdvo:*"
			],
			"Effect": "Allow"
		}
	]
}

Can you clarify which entries are redundant and should be removed?

[github-actions]

This issue has not received a response in a while. If you want to keep this issue open, please leave a comment below and auto-close will be canceled.

[github-actions]

⚠️COMMENT VISIBILITY WARNING⚠️

Comments on closed issues are hard for our team to see.
If you need more assistance, please either tag a team member or open a new issue that references this one.
If you wish to keep having a conversation with other community members under this issue feel free to do so.