安全审计>> iac-aws-iam>> 返回
项目作者： hmussana

项目描述 ：
IaC (Infrastructure as Code) IAM is a set of AWS CloudFormation templates for quickly setting up cross-account access.
高级语言：
项目地址: git://github.com/hmussana/iac-aws-iam.git
创建时间： 2019-10-07T19:46:39Z
项目社区：https://github.com/hmussana/iac-aws-iam
开源协议：The Unlicense
下载


iac-aws-iam

Introduction

IaC AWS IAM is a set of example AWS CloudFormation templates for quickly setting up infrastructure pipelines for Cross-Account Access using AWS CodePipeline and GitHub repo as the pipeline trigger/source. Using this setup, IAM users(human users intended) only have to be created once in the jump account(manager account) and then can assume roles to other AWS accounts owned by the same owner(think AWS Organizations).
This repository contains cloudformation templates for setting up: IAM Users, IAM Groups and IAM Roles for those who:

  • have multiple AWS accounts and want to manage & roll out all human users automatically (Continuous deployment).
  • want to setup multiple AWS accounts from scratch and looking to implement cloudformation templates for AWS IAM users, roles and groups.

Services

  • AWS IAM: Identity and Access Management service for accessing AWS services
  • AWS CloudFormation: Infrastructure as Code Service for creating AWS resources. Yaml as the templates language
  • AWS CodePipeline: Continuous Delivery Service by AWS, used here to automatically deploy any changes pushed to the source git repository.

Usage

Terminology

  • manager-account: Main AWS account which only has IAM Groups and Users and acts as a jump account.
  • managed-accounts: Team/Project & environment wise AWS accounts e.g. development-staging or androidApp-production.
  • jump account: Another name for manager account.

Login Workflow

User Login: Login to manager account -> Assume role to a managed account based upon group membership

Directory Structure

  1. iac-aws-iam
  2. │
  3. ├── README.md
  4. ├── managed-accounts
  5. │   └── project1-production
  6. │       ├── pipeline.yaml
  7. │       └── roles.yaml
  8. └── manager-account
  9.     ├── groups-admin.yaml
  10.     ├── pipeline.yaml
  11.     └── users.yaml

Setup Instructions

Following two types of pipelines will be setup using only the already defined resources in templates:

  1. IAM Roles Pipeline: Should be setup first in the managed account(s) so the IAM groups referencing to AssumeRole of the relevant role will not point to a non-existent resource. Each AWS account should have its own pipeline for setting up roles.
  2. IAM Users and Groups Pipeline: Has to be setup once in the main/manager account.

Deployment Steps

  1. Rename project1-production as desired and add all the aws accounts as directories under: managed-accounts
    • Also adjust the TemplatePath values accordingly in pipeline.yaml.
  2. Replace the example AWS account ID 012345678901 in project1-production and subsequently for all managed accounts you want to use.
  3. Create the pipeline stack in AWS CloudFormation console or via api/cli using managed-accounts/project1-production/pipeline.yaml which will setup the roles stack itself. See this for how to do that.
  4. Provide your GitHub token for the parameter value. Check this guide for details on how to generate one.
  5. Update/add the example account ID mappings in groups-admin.yaml. Create more group templates if desired using this file and name appropriately e.g. groups-developer.yaml or groups-readonly.yaml. Also change project names as desired and adjust GroupName and Output Export values accordingly.
  6. Add more list items under Action under Group-Stage in the pipeline with correct TemplatePath for each new template added(assuming more group templates were created in above step).
  7. Add more IAM users in the users.yaml as desired and adjust passwords and group memberships.
    • Important The GroupName attribute is being imported from the IAM groups stack so the names must match.
  8. Login to manager account and create the pipeline stack using manager-account/pipeline.yaml and provide GuitHub access token again.

Once the pipelines finish, the users should be able to login using their one-time passwords and can then assume roles into various AWS accounts as defined by the IAM groups they are members of.

Author

hmussana

References

Here are some documentation links which can be used as references:

  1. https://docs.aws.amazon.com/AWSCloudFormation/latest/UserGuide/aws-resource-codepipeline-pipeline.html
  2. https://docs.aws.amazon.com/AWSCloudFormation/latest/UserGuide/best-practices.html
  3. https://docs.aws.amazon.com/codepipeline/latest/userguidereference-pipeline-structure.html#pipeline-requirements