IAM policies (JSON structure) in AWS - Commands & Configuration

Introduction

IAM policies control who can do what in your AWS account. They are written in JSON and tell AWS which actions are allowed or denied on which resources.

When you want to give a user permission to access only specific AWS services.

When you need to restrict access to certain parts of your AWS resources for security.

When you want to allow an application to perform actions on AWS on your behalf.

When you want to create reusable permission sets for groups of users.

When you want to audit and control access to your AWS environment.

Config File - example-policy.json

example-policy.json

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "s3:ListBucket",
        "s3:GetObject"
      ],
      "Resource": [
        "arn:aws:s3:::example-bucket",
        "arn:aws:s3:::example-bucket/*"
      ]
    }
  ]
}

Version: The policy language version.

Statement: A list of permissions.

Effect: Whether to allow or deny the actions.

Action: The AWS actions this policy allows or denies.

Resource: The specific AWS resources this policy applies to.

Commands

This command creates a new IAM policy named ExampleS3ReadOnlyPolicy using the JSON file example-policy.json. It defines permissions for S3 read-only access to a specific bucket.

Terminal

aws iam create-policy --policy-name ExampleS3ReadOnlyPolicy --policy-document file://example-policy.json

Expected Output

{

    "Policy": {

        "PolicyName": "ExampleS3ReadOnlyPolicy",

        "PolicyId": "ABCDEFGHIJKLMN1234567",

        "Arn": "arn:aws:iam::123456789012:policy/ExampleS3ReadOnlyPolicy",

        "Path": "/",

        "DefaultVersionId": "v1",

        "AttachmentCount": 0,

        "PermissionsBoundaryUsageCount": 0,

        "IsAttachable": true,

        "CreateDate": "2024-06-01T12:00:00Z",

        "UpdateDate": "2024-06-01T12:00:00Z"

    }

}

→

--policy-name - Sets the name of the new IAM policy.

→

--policy-document - Specifies the JSON file that contains the policy.

This command lists all customer-created IAM policies in your AWS account to verify the new policy was created.

Terminal

aws iam list-policies --scope Local

Expected Output

Policies:

- ExampleS3ReadOnlyPolicy

- AnotherCustomPolicy

- MyTestPolicy

→

--scope - Filters policies to only those created in your account (not AWS managed).

Key Concept

If you remember nothing else from this pattern, remember: IAM policies are JSON documents that explicitly allow or deny actions on AWS resources.

Common Mistakes

Using incorrect JSON syntax in the policy document.

AWS will reject the policy because it must be valid JSON.

Always validate your JSON syntax before applying the policy.

Not specifying the correct resource ARN in the policy.

The policy will not apply to the intended resources, causing permission errors.

Use the exact ARN format for the AWS resource you want to control.

Forgetting to set the Effect to Allow or Deny.

The policy will be invalid without an Effect, so AWS will reject it.

Always include Effect with either "Allow" or "Deny" in each statement.

Summary

Create IAM policies as JSON files defining allowed or denied actions on AWS resources.

Use the AWS CLI to create and list policies to manage permissions.

Always validate JSON syntax and specify correct resource ARNs to avoid errors.

Practice

1. What is the main purpose of an IAM policy in AWS?

easy

A. To create virtual machines

B. To define permissions for users and resources

C. To monitor network traffic

D. To store data in the cloud

Solution

1. Step 1: Understand IAM policy role

   An IAM policy is a JSON document that specifies permissions for AWS users, groups, or roles.

2. Step 2: Identify main function

   Its main function is to control what actions are allowed or denied on AWS resources.

3. Final Answer:

   To define permissions for users and resources -> Option B

4. Quick Check:

   IAM policy = permissions definition [OK]

Hint: IAM policies control access permissions in AWS [OK]

Common Mistakes:

• Confusing IAM policies with data storage
• Thinking IAM policies monitor network traffic
• Assuming IAM policies create virtual machines

2. Which of the following is the correct JSON key to specify the effect of a statement in an IAM policy?

easy

A. "Permission"

B. "Action"

C. "Resource"

D. "Effect"

Solution

1. Step 1: Recall IAM policy statement keys

   IAM policy statements include keys like Effect, Action, Resource, and optionally Condition.

2. Step 2: Identify key for permission type

   The key that specifies whether to allow or deny is "Effect".

3. Final Answer:

   "Effect" -> Option D

4. Quick Check:

   Effect key = permission type [OK]

Hint: Effect key sets allow or deny in IAM policy [OK]

Common Mistakes:

• Using "Permission" instead of "Effect"
• Confusing "Action" with permission type
• Mistaking "Resource" for effect

3. Given this IAM policy statement snippet:

{
  "Effect": "Allow",
  "Action": "s3:ListBucket",
  "Resource": "arn:aws:s3:::example-bucket"
}

What permission does this statement grant?

medium

A. Allows listing the bucket itself

B. Allows listing objects inside the bucket

C. Allows deleting the bucket

D. Allows uploading objects to the bucket

Solution

1. Step 1: Understand the Action "s3:ListBucket"

   This action allows listing the bucket itself and its metadata, not the objects inside.

2. Step 2: Match Resource and Action

   The resource is the bucket ARN, so permission is to list the bucket (its properties), not the objects inside the bucket.

3. Final Answer:

   Allows listing the bucket itself -> Option A

4. Quick Check:

   s3:ListBucket = list bucket (not objects) [OK]

Hint: s3:ListBucket lists the bucket, not objects inside [OK]

Common Mistakes:

• Confusing ListBucket with listing objects inside the bucket
• Assuming permission to delete or upload
• Ignoring the resource ARN level

4. Identify the error in this IAM policy statement:

{
  "Effect": "Allow",
  "Action": ["ec2:StartInstances", "ec2:StopInstances"],
  "Resource": "*",
  "Condition": {
    "StringEquals": {
      "ec2:Region": "us-west-2"
    }
  }
}

medium

A. The Condition key is not valid for EC2 actions

B. The Condition key should be inside the Action key

C. The policy is valid and has no errors

D. The Resource value "*" is not allowed for these actions

Solution

1. Step 1: Check Condition usage with EC2 actions

   EC2 supports conditions like StringEquals on ec2:Region to restrict actions by region.

2. Step 2: Verify Resource and structure

   Resource "*" is valid for EC2 start/stop actions because they apply to instances across resources.

3. Final Answer:

   The policy is valid and has no errors -> Option C

4. Quick Check:

   Condition on ec2:Region with Resource "*" is valid [OK]

Hint: Conditions can restrict actions by region or other keys [OK]

Common Mistakes:

• Thinking Condition is invalid for EC2
• Assuming Resource "*" is always wrong
• Misplacing Condition inside Action

5. You want to create an IAM policy that allows a user to read objects only from a specific S3 bucket named "my-data-bucket" but denies deleting any objects. Which policy statement correctly achieves this?

hard

A. {
"Effect": "Allow",
"Action": ["s3:GetObject"],
"Resource": "arn:aws:s3:::my-data-bucket/*"
}

B. {
"Effect": "Allow",
"Action": ["s3:GetObject", "s3:DeleteObject"],
"Resource": "arn:aws:s3:::my-data-bucket/*"
}

C. {
"Effect": "Deny",
"Action": "s3:DeleteObject",
"Resource": "arn:aws:s3:::my-data-bucket/*"
}

D. {
"Effect": "Allow",
"Action": "s3:*",
"Resource": "arn:aws:s3:::my-data-bucket"
}

Solution

1. Step 1: Identify required permissions

   The user needs permission to read objects only, which is "s3:GetObject" on the bucket's objects.

2. Step 2: Check for delete denial

   Not including "s3:DeleteObject" means no delete permission is granted. Explicit deny is not required if no allow exists.

3. Step 3: Validate resource ARN

   The resource must include "/*" to specify objects inside the bucket, not the bucket itself.

4. Final Answer:

   Allow s3:GetObject on objects in my-data-bucket only -> Option A

5. Quick Check:

   Allow read only, no delete = { "Effect": "Allow", "Action": ["s3:GetObject"], "Resource": "arn:aws:s3:::my-data-bucket/*" } [OK]

Hint: Allow only needed actions; omit delete to deny it [OK]

Common Mistakes:

• Allowing delete by mistake
• Using bucket ARN without /* for objects
• Using wildcard s3:* granting too many permissions