Feat/multi org support

README.md

@@ -38,3 +38,5 @@ template as its baseline.
   contribute or build on top of ADF.
 - Refer to the [Samples Guide](docs/samples-guide.md) for a detailed walk
   through of the provided samples.
+- Refer to the [Multi-Organization ADF Setup](docs/multi-organization-guide.md)
+   to use ADF in an enterprise-grade setup.

docs/admin-guide.md

@@ -100,6 +100,8 @@ config:
   scm: # Source Control Management
     auto-create-repositories: enabled # Optional
     default-scm-branch: main          # Optional
+  org:
+    stage: dev # Optional
 ```
 
 In the above example the properties are categorized into `roles`, `regions`,
@@ -232,10 +234,13 @@ Config has five components in `main-notification-endpoint`, `scp`, `scm`,
     and prod AWS Organization with its own ADF instance per AWS organization.
     This approach allows for well-tested and stable prod AWS Organization
     deployments.  If set, a matching SSM parameter `/adf/org/stage` gets
-    created that you can reference in your buildspec files to allow for
+    created in the deployment and all target accounts.
+    You can reference it within your buildspec files to allow for
     org-specific deployments; without hardcoding the AWS Organization stage in
     your buildspec. If this variable is not set, the SSM parameter
     `/adf/org/stage` defaults to "none".
+    More information about setting up ADF with multiple AWS Organizations can
+    be found in the [Multi-Organization Guide](multi-organization-guide.md)
 
   - **default-scm-codecommit-account-id** allows you to configure the default
     account id that should be used with all source-code management platforms

src/lambda_codebase/account_bootstrap.py

@@ -67,6 +67,9 @@ def configure_generic_account(sts, event, region, role):
         bucket_name = parameter_store_deployment_account.fetch_parameter(
             f'/cross_region/s3_regional_bucket/{region}',
         )
+        org_stage = parameter_store_deployment_account.fetch_parameter(
+            '/adf/org/stage'
+        )
     except (ClientError, ParameterNotFoundError):
         raise GenericAccountConfigureError(
             f'Account {event["account_id"]} cannot yet be bootstrapped '
@@ -79,6 +82,7 @@ def configure_generic_account(sts, event, region, role):
         'deployment_account_id',
         event['deployment_account_id'],
     )
+    parameter_store_target_account.put_parameter('/adf/org/stage', org_stage)
 
 
 def configure_master_account_parameters(event):

src/lambda_codebase/initial_commit/bootstrap_repository/adf-bootstrap/deployment/example-global-iam.yml

@@ -4,6 +4,28 @@
 AWSTemplateFormatVersion: "2010-09-09"
 Description: ADF CloudFormation Template (Global) for IAM in the Deployment Account
 
+# Parameters:
+#   OrgStage can be set in the respective adfconfig file using the
+#   path config.org.stage
+#   OrgStage:
+#     Type: "AWS::SSM::Parameter::Value<String>"
+#     Description: A stage used to differentiate Multi-Organization ADF environments
+#     Default: /adf/org/stage
+
+# Org StageCustom Mappings allows you to dynamically build different IAM
+# Conditions / Principals ARN / Resource ARN per Organization applying Least
+# Privilege Principals whilst retaining a Single Stack Definition for all
+# environments.
+# Usage: !FindInMap[OrgStageMap: !Ref OrgStage, ExampleCustomProperty]
+# Mappings:
+#   OrgStageMap:
+#     Dev:
+#       ExampleCustomProperty: 1234
+#     Int:
+#       ExampleCustomProperty: 5678
+#     Prod:
+#       ExampleCustomProperty: 9102
+
 Resources:
   CloudFormationDeploymentPolicy:
     # This is the policy that will be used to deploy CloudFormation resources from

src/lambda_codebase/initial_commit/bootstrap_repository/adf-bootstrap/example-global-iam.yml

@@ -9,6 +9,26 @@ Parameters:
     Type: "AWS::SSM::Parameter::Value<String>"
     Description: Deployment Account ID
     Default: deployment_account_id
+    # OrgStage can be set in the respective adfconfig file using the
+    # path config.org.stage
+    # OrgStage:
+    #   Type: "AWS::SSM::Parameter::Value<String>"
+    #   Description: A stage used to differentiate Multi-Org ADF environments
+    #   Default: /adf/org/stage
+
+# Org StageCustom Mappings allows you to dynamically build different IAM
+# Conditions / Principals ARN / Resource ARN per Organization applying
+# least-privilege principles whilst retaining a Single Stack Definition for all
+# environments.
+# usage !FindInMap[OrgStageMap: !Ref OrgStage, ExampleCustomProperty]
+# Mappings:
+#   OrgStageMap:
+#     Dev:
+#       ExampleCustomProperty: 1234
+#     Int:
+#       ExampleCustomProperty: 5678
+#     Prod:
+#       ExampleCustomProperty: 9102
 
 Resources:
   CloudFormationDeploymentPolicy:

src/lambda_codebase/initial_commit/bootstrap_repository/adf-build/config.py

@@ -90,11 +90,20 @@ def _validate(self):
 
     def _load_config_file(self):
         """
-        Loads the adfconfig.yml file and executes _parse_config
+        Checks for an Org Specific adfconfig.yml (adfconfig.{ORG_ID}.yml)
+        and uses that if it exists. Otherwise it uses the default adfconfig.yml
+        and executes _parse_config
         """
-        with open(self.config_path, encoding="utf-8") as config:
-            self.config_contents = yaml.load(config, Loader=yaml.FullLoader)
-            self._parse_config()
+        org_config_path = self.config_path.replace(".yml", f".{self.organization_id}.yml")
+        if os.path.exists(org_config_path):
+            with open(org_config_path, encoding="utf-8") as org_config_file:
+                LOGGER.info("Using custom org config")
+                self.config_contents = yaml.load(org_config_file, Loader=yaml.FullLoader)
+        else:
+            LOGGER.info("Using default org config")
+            with open(self.config_path, encoding="utf-8") as config:
+                self.config_contents = yaml.load(config, Loader=yaml.FullLoader)
+        self._parse_config()
 
     def _parse_config(self):
         """

src/lambda_codebase/initial_commit/bootstrap_repository/adf-build/main.py

@@ -270,6 +270,15 @@ def worker_thread(
                 'bucket_name',
                 updated_kms_bucket_dict[region]['s3_regional_bucket'],
             )
+
+            # Ensuring the stage parameter on the target account is up-to-date
+            parameter_store.put_parameter(
+                '/adf/org/stage',
+                config.config.get('org', {}).get(
+                    'stage',
+                    ADF_DEFAULT_ORG_STAGE,
+                )
+            )
             cloudformation = CloudFormation(
                 region=region,
                 deployment_account_region=config.deployment_account_region,

src/lambda_codebase/initial_commit/bootstrap_repository/example-adfconfig.yml

@@ -28,6 +28,13 @@ config:
     default-scm-branch: master
     # ^ The default branch is used when the pipeline does not specify a specific branch.
     # If this parameter is not specified, it defaults to the "master" branch.
+  org:
+    stage: prod
+    # ^ This value will be set as an SSM Parameter named /adf/org/stage
+    # in both the deployment account and and all
+    # Target member accounts as part of the Bootstrap Statemachine.
+    # It is useful as a flag to drive Organization specific logic within
+    # IAM Role definitions/conditions as well as Build spec behavior.
 
 extensions:
   terraform: