CloudPath Academy

Your guide to AWS certification success

Amazon Web Services AWS Broken Labs

Lab 01 - Access Denied I

Difficulty: Beginner Service: AWS IAM, Amazon S3

Cost: This lab deploys only IAM resources — no charges apply.

Scenario

An application team was granted an IAM role so their service can read objects from an S3 bucket. The stack deployed without errors and the role has a policy attached — but the team reports that read requests are still being denied.

What Was Deployed

Resource Purpose
AWS::S3::Bucket Target bucket the role is meant to read from
AWS::IAM::Role Role the application uses to read from the bucket

The stack deployed without errors. The role exists and has a policy attached.

Deploy the Lab

  1. Open the AWS CloudFormation console
  2. Click Create stack > With new resources (standard)
  3. Select Upload a template file and upload lab-01-access-denied-i.yaml
  4. Enter a stack name (e.g., brokenlabs-iam-lab-01), click Next > Next
  5. Check I acknowledge that AWS CloudFormation might create IAM resources and click Submit
  6. Wait for the stack status to reach CREATE_COMPLETE
  7. Open the stack Outputs tab — note RoleName, BucketResourceArn, and ObjectResourceArn

The Problem

Open the IAM Policy Simulator, select the role named in the stack Outputs (RoleName), choose the s3:GetObject action, and set the resource ARN to the ObjectResourceArn value from Outputs. Run the simulation.

Note: If you’re using the Simulate button inside the IAM console’s policy editor instead of the standalone Policy Simulator, each action you add gets its own Resource field that defaults to *. Enter the ARN from the stack Outputs instead of leaving the default *, which can show denied regardless of the policy.

Expected: the simulation result for s3:GetObject shows allowed. Actual: the simulation result shows denied.

Fix the Lab

Diagnose why the simulation is denied and fix the role’s permissions so the result matches Expected above. Re-run the simulation to confirm your fix.

Need help? Open hints.md for progressive hints.

Cleanup

  1. Open CloudFormation, select your stack, and click Delete

Resources