[sdlf-dataset] handling of storage parameters

aws-solutions-library-samples · Nov 25, 2024 · e1b5757 · e1b5757
1 parent c628976
commit e1b5757
Showing 1 changed file with 67 additions and 43 deletions.
diff --git a/sdlf-dataset/src/dataset.yaml b/sdlf-dataset/src/dataset.yaml
@@ -9,37 +9,42 @@ Parameters:
     Type: String
     Description: A string uniquely identifying this deployment in this AWS account
     Default: dev
+  pStorageDeploymentInstance:
+    Type: String
+    Description: The string uniquely identifying a sdlf-foundations deployment in this AWS account
+    Default: "" # see below comments
   pOrg:
     Description: Name of the organization owning the datalake
     Type: String
-    Default: "{{resolve:ssm:/sdlf/storage/rOrganization/dev}}"
+    Default: "" # if not provided, pStorageDeploymentInstance must be specified
   pDomain:
     Description: Data domain name
     Type: String
-    Default: "{{resolve:ssm:/sdlf/storage/rDomain/dev}}"
+    Default: "" # if not provided, pStorageDeploymentInstance must be specified
+  pBucketKey:
+    Description: KMS key set as bucket key for the solutions' buckets
+    Type: String
+    Default: "" # if not provided, pStorageDeploymentInstance must be specified
   pRawBucket:
-    Description: The raw bucket for the solution
+    Description: Raw bucket
     Type: String
-    Default: "{{resolve:ssm:/sdlf/storage/rRawBucket/dev}}"
+    Default: "" # if not provided, pStorageDeploymentInstance must be specified
   pStageBucket:
-    Description: The stage bucket for the solution
+    Description: Stage bucket
     Type: String
-    Default: "{{resolve:ssm:/sdlf/storage/rStageBucket/dev}}"
+    Default: "" # if not provided, pStorageDeploymentInstance must be specified
   pAnalyticsBucket:
-    Description: The analytics bucket for the solution
+    Description: Analytics bucket
     Type: String
-    Default: "{{resolve:ssm:/sdlf/storage/rAnalyticsBucket/dev}}"
+    Default: "" # if not provided, pStorageDeploymentInstance must be specified
   pArtifactsBucket:
     Description: S3 bucket used to store artifacts (from CICD or generated by data pipelines)
     Type: String
-    Default: "{{resolve:ssm:/sdlf/storage/rArtifactsBucket/dev}}"
-  pBucketKey:
-    Description: KMS key set as bucket key for the solutions' buckets
-    Type: String
-    Default: "{{resolve:ssm:/sdlf/storage/rKMSKey/dev}}"
+    Default: "" # if not provided, pStorageDeploymentInstance must be specified
   pLakeFormationDataAccessRole:
+    Description: https://docs.aws.amazon.com/lake-formation/latest/dg/registration-role.html
     Type: String
-    Default: "{{resolve:ssm:/sdlf/storage/rLakeFormationDataAccessRoleArn/dev}}"
+    Default: "" # if not provided, pStorageDeploymentInstance must be specified
   pDatasetName:
     Description: The name of the dataset (all lowercase, no symbols or spaces)
     Type: String
@@ -59,6 +64,7 @@ Parameters:
     Default: false
 
 Conditions:
+  FetchFromStorageSsm: !Not [!Equals [!Ref pStorageDeploymentInstance, ""]]
   IsS3Prefix: !Not [!Equals [!Ref pS3Prefix, ""]]
   RunInVpc: !Equals [!Ref pEnableVpc, true]
 
@@ -165,7 +171,7 @@ Resources:
               - kms:Describe*
             Effect: Allow
             Principal:
-              AWS: !Ref pLakeFormationDataAccessRole
+              AWS: !If [FetchFromStorageSsm, !Sub "{{resolve:ssm:/sdlf/storage/rLakeFormationDataAccessRoleArn/${pStorageDeploymentInstance}}}", !Ref pLakeFormationDataAccessRole]
             Resource: "*"
 
   rKMSDataKeyAlias:
@@ -203,7 +209,7 @@ Resources:
             KmsKeyArn: !If
               - IsS3Prefix
               - !GetAtt rKMSDataKey.Arn
-              - !Ref pBucketKey
+              - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rKMSKey/${pStorageDeploymentInstance}}}", !Ref pBucketKey]
 
   rGlueSecurityConfigurationSsm:
     Type: AWS::SSM::Parameter
@@ -326,9 +332,9 @@ Resources:
                   - s3:PutObject
                   - s3:PutObjectVersion
                 Resource:
-                  - !Sub arn:${AWS::Partition}:s3:::${pRawBucket}/${pS3Prefix}/*
-                  - !Sub arn:${AWS::Partition}:s3:::${pStageBucket}/${pS3Prefix}/*
-                  - !Sub arn:${AWS::Partition}:s3:::${pAnalyticsBucket}/${pS3Prefix}/*
+                  - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rRawBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}/*", !Sub "arn:${AWS::Partition}:s3:::${pRawBucket}/${pS3Prefix}/*"]
+                  - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rStageBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}/*", !Sub "arn:${AWS::Partition}:s3:::${pStageBucket}/${pS3Prefix}/*"]
+                  - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rAnalyticsBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}/*", !Sub "arn:${AWS::Partition}:s3:::${pAnalyticsBucket}/${pS3Prefix}/*"]
               - Effect: Allow
                 Action:
                   - kms:DescribeKey
@@ -341,9 +347,9 @@ Resources:
                   - IsS3Prefix
                   - - !GetAtt rKMSInfraKey.Arn
                     - !GetAtt rKMSDataKey.Arn
-                    - !Ref pBucketKey
+                    - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rKMSKey/${pStorageDeploymentInstance}}}", !Ref pBucketKey]
                   - - !GetAtt rKMSInfraKey.Arn
-                    - !Ref pBucketKey
+                    - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rKMSKey/${pStorageDeploymentInstance}}}", !Ref pBucketKey]
               - Effect: Allow
                 Action:
                   - lakeformation:GetDataAccess # W11 exception
@@ -379,7 +385,10 @@ Resources:
       CatalogId: !Ref AWS::AccountId
       DatabaseInput:
         Description: !Sub "${pDatasetName} raw metadata catalog"
-        Name: !Sub ${pOrg}_${pDomain}_${pDatasetName}_raw
+        Name: !If
+          - FetchFromStorageSsm
+          - !Sub "{{resolve:ssm:/sdlf/storage/rOrganization/${pStorageDeploymentInstance}}}_{{resolve:ssm:/sdlf/storage/rDomain/${pStorageDeploymentInstance}}}_${pDatasetName}_raw"
+          - !Sub ${pOrg}_${pDomain}_${pDatasetName}_raw
 
   rRawGlueDataCatalogSsm:
     Type: AWS::SSM::Parameter
@@ -411,7 +420,7 @@ Resources:
       Name: !Sub sdlf-${pDatasetName}-raw-crawler
       Targets:
         S3Targets:
-          - Path: !Sub s3://${pRawBucket}/${pS3Prefix}
+          - Path: !If [FetchFromStorageSsm, !Sub "s3://{{resolve:ssm:/sdlf/storage/rRawBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}", !Sub "s3://${pRawBucket}/${pS3Prefix}"]
 
   rRawGlueCrawlerGlueLakeFormationPermissions:
     Type: AWS::LakeFormation::Permissions
@@ -435,7 +444,7 @@ Resources:
         - DATA_LOCATION_ACCESS
       Resource:
         DataLocationResource:
-          S3Resource: !Sub arn:${AWS::Partition}:s3:::${pRawBucket}/${pS3Prefix}/
+          S3Resource: !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rRawBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}/", !Sub "arn:${AWS::Partition}:s3:::${pRawBucket}/${pS3Prefix}/"]
 
   rRawGlueCrawlerSsm:
     Type: AWS::SSM::Parameter
@@ -451,7 +460,10 @@ Resources:
       CatalogId: !Ref AWS::AccountId
       DatabaseInput:
         Description: !Sub "${pDatasetName} stage metadata catalog"
-        Name: !Sub ${pOrg}_${pDomain}_${pDatasetName}_stage
+        Name: !If
+          - FetchFromStorageSsm
+          - !Sub "{{resolve:ssm:/sdlf/storage/rOrganization/${pStorageDeploymentInstance}}}_{{resolve:ssm:/sdlf/storage/rDomain/${pStorageDeploymentInstance}}}_${pDatasetName}_stage"
+          - !Sub ${pOrg}_${pDomain}_${pDatasetName}_stage
 
   rStageGlueDataCatalogSsm:
     Type: AWS::SSM::Parameter
@@ -483,7 +495,7 @@ Resources:
       Name: !Sub sdlf-${pDatasetName}-stage-crawler
       Targets:
         S3Targets:
-          - Path: !Sub s3://${pStageBucket}/${pS3Prefix}
+          - Path: !If [FetchFromStorageSsm, !Sub "s3://{{resolve:ssm:/sdlf/storage/rStageBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}", !Sub "s3://${pStageBucket}/${pS3Prefix}"]
 
   rStageGlueCrawlerGlueLakeFormationPermissions:
     Type: AWS::LakeFormation::Permissions
@@ -507,7 +519,7 @@ Resources:
         - DATA_LOCATION_ACCESS
       Resource:
         DataLocationResource:
-          S3Resource: !Sub arn:${AWS::Partition}:s3:::${pStageBucket}/${pS3Prefix}/
+          S3Resource: !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rStageBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}/", !Sub "arn:${AWS::Partition}:s3:::${pStageBucket}/${pS3Prefix}/"]
 
   rStageGlueCrawlerSsm:
     Type: AWS::SSM::Parameter
@@ -523,7 +535,10 @@ Resources:
       CatalogId: !Ref AWS::AccountId
       DatabaseInput:
         Description: !Sub "${pDatasetName} analytics metadata catalog"
-        Name: !Sub ${pOrg}_${pDomain}_${pDatasetName}_analytics
+        Name: !If
+          - FetchFromStorageSsm
+          - !Sub "{{resolve:ssm:/sdlf/storage/rOrganization/${pStorageDeploymentInstance}}}_{{resolve:ssm:/sdlf/storage/rDomain/${pStorageDeploymentInstance}}}_${pDatasetName}_analytics"
+          - !Sub ${pOrg}_${pDomain}_${pDatasetName}_analytics
 
   rAnalyticsGlueDataCatalogSsm:
     Type: AWS::SSM::Parameter
@@ -555,7 +570,7 @@ Resources:
       Name: !Sub sdlf-${pDatasetName}-analytics-crawler
       Targets:
         S3Targets:
-          - Path: !Sub s3://${pAnalyticsBucket}/${pS3Prefix}
+          - Path: !If [FetchFromStorageSsm, !Sub "s3://{{resolve:ssm:/sdlf/storage/rAnalyticsBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}", !Sub "s3://${pAnalyticsBucket}/${pS3Prefix}"]
 
   rAnalyticsGlueCrawlerGlueLakeFormationPermissions:
     Type: AWS::LakeFormation::Permissions
@@ -579,7 +594,7 @@ Resources:
         - DATA_LOCATION_ACCESS
       Resource:
         DataLocationResource:
-          S3Resource: !Sub arn:${AWS::Partition}:s3:::${pAnalyticsBucket}/${pS3Prefix}/
+          S3Resource: !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rAnalyticsBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}/", !Sub "arn:${AWS::Partition}:s3:::${pAnalyticsBucket}/${pS3Prefix}/"]
 
   rAnalyticsGlueCrawlerSsm:
     Type: AWS::SSM::Parameter
@@ -701,21 +716,21 @@ Resources:
             Action:
               - s3:ListBucket
             Resource:
-              - !Sub arn:${AWS::Partition}:s3:::${pArtifactsBucket}
-              - !Sub arn:${AWS::Partition}:s3:::${pRawBucket}
-              - !Sub arn:${AWS::Partition}:s3:::${pStageBucket}
-              - !Sub arn:${AWS::Partition}:s3:::${pAnalyticsBucket}
+              - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rArtifactsBucket/${pStorageDeploymentInstance}}}", !Sub "arn:${AWS::Partition}:s3:::${pArtifactsBucket}"]
+              - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rRawBucket/${pStorageDeploymentInstance}}}", !Sub "arn:${AWS::Partition}:s3:::${pRawBucket}"]
+              - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rStageBucket/${pStorageDeploymentInstance}}}", !Sub "arn:${AWS::Partition}:s3:::${pStageBucket}"]
+              - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rAnalyticsBucket/${pStorageDeploymentInstance}}}", !Sub "arn:${AWS::Partition}:s3:::${pAnalyticsBucket}"]
           - Sid: AllowTeamPrefixActions
             Effect: Allow
             Action:
               - s3:DeleteObject
               - s3:GetObject
               - s3:PutObject
             Resource:
-              - !Sub arn:${AWS::Partition}:s3:::${pArtifactsBucket}/${pS3Prefix}/*
-              - !Sub arn:${AWS::Partition}:s3:::${pRawBucket}/${pS3Prefix}/*
-              - !Sub arn:${AWS::Partition}:s3:::${pStageBucket}/${pS3Prefix}/*
-              - !Sub arn:${AWS::Partition}:s3:::${pAnalyticsBucket}/${pS3Prefix}/*
+              - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rArtifactsBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}/*", !Sub "arn:${AWS::Partition}:s3:::${pArtifactsBucket}/${pS3Prefix}/*"]
+              - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rRawBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}/*", !Sub "arn:${AWS::Partition}:s3:::${pRawBucket}/${pS3Prefix}/*"]
+              - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rStageBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}/*", !Sub "arn:${AWS::Partition}:s3:::${pStageBucket}/${pS3Prefix}/*"]
+              - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rAnalyticsBucket/${pStorageDeploymentInstance}}}/${pS3Prefix}/*", !Sub "arn:${AWS::Partition}:s3:::${pAnalyticsBucket}/${pS3Prefix}/*"]
           - Sid: AllowTeamKMSDataKeyUsage
             Effect: Allow
             Action:
@@ -729,9 +744,9 @@ Resources:
               - IsS3Prefix
               - - !GetAtt rKMSInfraKey.Arn
                 - !GetAtt rKMSDataKey.Arn
-                - !Ref pBucketKey
+                - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rKMSKey/${pStorageDeploymentInstance}}}", !Ref pBucketKey]
               - - !GetAtt rKMSInfraKey.Arn
-                - !Ref pBucketKey
+                - !If [FetchFromStorageSsm, !Sub "arn:${AWS::Partition}:s3:::{{resolve:ssm:/sdlf/storage/rKMSKey/${pStorageDeploymentInstance}}}", !Ref pBucketKey]
           - Effect: Allow
             Action:
               - ssm:GetParameter
@@ -793,11 +808,20 @@ Resources:
               - glue:GetDataQualityRule*
             Resource:
               - !Sub arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:catalog
-              - !Sub arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:database/${pOrg}_${pDomain}_${pDatasetName}_*
-              - !Sub arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:table/${pOrg}_${pDomain}_${pDatasetName}_*
+              - !If
+                - FetchFromStorageSsm
+                - !Sub "arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:database/{{resolve:ssm:/sdlf/storage/rOrganization/${pStorageDeploymentInstance}}}_{{resolve:ssm:/sdlf/storage/rDomain/${pStorageDeploymentInstance}}}_${pDatasetName}_*"
+                - !Sub arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:database/${pOrg}_${pDomain}_${pDatasetName}_*
+              - !If
+                - FetchFromStorageSsm
+                - !Sub "arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:table/{{resolve:ssm:/sdlf/storage/rOrganization/${pStorageDeploymentInstance}}}_{{resolve:ssm:/sdlf/storage/rDomain/${pStorageDeploymentInstance}}}_${pDatasetName}_*"
+                - !Sub arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:table/${pOrg}_${pDomain}_${pDatasetName}_*
+              - !If
+                - FetchFromStorageSsm
+                - !Sub "arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:job/{{resolve:ssm:/sdlf/storage/rOrganization/${pStorageDeploymentInstance}}}-{{resolve:ssm:/sdlf/storage/rDomain/${pStorageDeploymentInstance}}}-${pDatasetName}_*"
+                - !Sub arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:job/${pOrg}-${pDomain}-${pDatasetName}-*
               - !Sub arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:crawler/sdlf-${pDatasetName}-*
               - !Sub arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:job/sdlf-${pDatasetName}-*
-              - !Sub arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:job/${pOrg}-${pDomain}-${pDatasetName}-*
               - !Sub arn:${AWS::Partition}:glue:${AWS::Region}:${AWS::AccountId}:dataQualityRuleset/* # glue:StartDataQualityRuleRecommendationRun requires dataQualityRuleset/*
           - Effect: Allow
             Action: