本文為英文版的機器翻譯版本，如內容有任何歧義或不一致之處，概以英文版為準。

# IAM 角色設定
<a name="spark-troubleshooting-agent-iam-setup"></a>

## 先決條件
<a name="spark-troubleshooting-agent-iam-prerequisites"></a>

開始前，請確保您具備以下條件：
+ 具有 IAM 管理存取權 AWS 的帳戶
+ AWS 已安裝並設定 CLI。如需詳細資訊，請參閱[安裝 AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html)。

設定下列變數以用於後續命令：

```
ACCOUNT_ID=$(aws sts get-caller-identity --query Account --output text)
REGION=$(aws configure get region)
```

## 步驟 1：建立 IAM 角色
<a name="spark-troubleshooting-agent-iam-step1"></a>

SMUS MCP 伺服器使用您的 IAM 角色來授權 AWS 服務層級的操作。不需要單獨的 MCP 特定許可。

**建立 IAM 角色 (AWS CLI)**

1. 建立信任政策文件，允許您的帳戶擔任該角色：

   ```
   cat > mcp-trust-policy.json << EOF
   {
     "Version": "2012-10-17",
     "Statement": [
       {
         "Sid": "AllowAccountToAssumeRole",
         "Effect": "Allow",
         "Principal": { "AWS": "arn:aws:iam::${ACCOUNT_ID}:root" },
         "Action": "sts:AssumeRole"
       }
     ]
   }
   EOF
   ```

1. 建立角色：

   ```
   aws iam create-role \
     --role-name SparkTroubleshootingMCPRole \
     --assume-role-policy-document file://mcp-trust-policy.json
   ```

## 步驟 2：連接部署模式的許可
<a name="spark-troubleshooting-agent-iam-step2"></a>

連接符合您 Spark 部署平台的許可政策。視您使用的平台而定，您可以連接下列一或多個 。

### 選項 A：EC2 上的 EMR
<a name="spark-troubleshooting-agent-iam-emr-ec2"></a>

1. 建立政策文件：

   ```
   cat > emr-ec2-policy.json << 'EOF'
   {
     "Version": "2012-10-17",		 	 	 
     "Statement": [
       {
         "Sid": "EMREC2ReadAccess",
         "Effect": "Allow",
         "Action": [
           "elasticmapreduce:DescribeCluster",
           "elasticmapreduce:DescribeStep",
           "elasticmapreduce:ListSteps",
           "elasticmapreduce:ListClusters",
           "elasticmapreduce:DescribeJobFlows"
         ],
         "Resource": ["*"]
       },
       {
         "Sid": "EMRS3LogAccess",
         "Effect": "Allow",
         "Action": ["s3:GetObject", "s3:ListBucket"],
         "Resource": "*"
       },
       {
         "Sid": "EMRPersistentApp",
         "Effect": "Allow",
         "Action": [
           "elasticmapreduce:CreatePersistentAppUI",
           "elasticmapreduce:DescribePersistentAppUI",
           "elasticmapreduce:GetPersistentAppUIPresignedURL"
         ],
         "Resource": ["*"]
       }
     ]
   }
   EOF
   ```

1. 建立並連接政策：

   ```
   aws iam put-role-policy \
     --role-name SparkTroubleshootingMCPRole \
     --policy-name EMREC2TroubleshootingAccess \
     --policy-document file://emr-ec2-policy.json
   ```

或者，如果您的角色已使用 [AmazonElasticMapReduceFullAccess](https://docs.aws.amazon.com/aws-managed-policy/latest/reference/AmazonElasticMapReduceFullAccess.html) AWS 受管政策，您可以連接它：

```
aws iam attach-role-policy \
  --role-name SparkTroubleshootingMCPRole \
  --policy-arn arn:aws:iam::aws:policy/AmazonElasticMapReduceFullAccess
```

### 選項 B： AWS Glue
<a name="spark-troubleshooting-agent-iam-glue"></a>

1. 建立政策文件：

   ```
   cat > glue-policy.json << EOF
   {
     "Version": "2012-10-17",		 	 	 
     "Statement": [
       {
         "Sid": "GlueReadAccess",
         "Effect": "Allow",
         "Action": [
           "glue:GetJob",
           "glue:GetJobRun",
           "glue:GetJobRuns",
           "glue:GetJobs",
           "glue:BatchGetJobs"
         ],
         "Resource": ["arn:aws:glue:*:${ACCOUNT_ID}:job/*"]
       },
       {
         "Sid": "GlueCloudWatchLogsAccess",
         "Effect": "Allow",
         "Action": ["logs:GetLogEvents", "logs:FilterLogEvents"],
         "Resource": ["arn:aws:logs:*:${ACCOUNT_ID}:log-group:/aws/glue/*"]
       },
       {
         "Sid": "GlueSparkWebUI",
         "Effect": "Allow",
         "Action": [
           "glue:RequestLogParsing",
           "glue:GetLogParsingStatus",
           "glue:GetEnvironment",
           "glue:GetStage",
           "glue:GetStages",
           "glue:GetStageFiles",
           "glue:BatchGetStageFiles",
           "glue:GetStageAttempt",
           "glue:GetStageAttemptTaskList",
           "glue:GetStageAttemptTaskSummary",
           "glue:GetExecutors",
           "glue:GetExecutorsThreads",
           "glue:GetStorage",
           "glue:GetStorageUnit",
           "glue:GetQueries",
           "glue:GetQuery",
           "glue:GetDashboardUrl"
         ],
         "Resource": ["arn:aws:glue:*:${ACCOUNT_ID}:job/*"]
       },
       {
         "Sid": "GluePassRoleAccess",
         "Effect": "Allow",
         "Action": "iam:PassRole",
         "Resource": "*",
         "Condition": {
           "StringLike": {
             "iam:PassedToService": "glue.amazonaws.com"
           }
         }
       }
     ]
   }
   EOF
   ```

1. 連接政策：

   ```
   aws iam put-role-policy \
     --role-name SparkTroubleshootingMCPRole \
     --policy-name GlueTroubleshootingAccess \
     --policy-document file://glue-policy.json
   ```

### 選項 C：EMR Serverless
<a name="spark-troubleshooting-agent-iam-emr-serverless"></a>

1. 建立政策文件：

   ```
   cat > emr-serverless-policy.json << EOF
   {
     "Version": "2012-10-17",		 	 	 
     "Statement": [
       {
         "Sid": "EMRServerlessReadAccess",
         "Effect": "Allow",
         "Action": [
           "emr-serverless:GetJobRun",
           "emr-serverless:GetApplication",
           "emr-serverless:ListApplications",
           "emr-serverless:ListJobRuns",
           "emr-serverless:ListJobRunAttempts",
           "emr-serverless:GetDashboardForJobRun",
           "emr-serverless:ListTagsForResource"
         ],
         "Resource": ["*"]
       },
       {
         "Sid": "EMRServerlessCloudWatchLogsAccess",
         "Effect": "Allow",
         "Action": ["logs:GetLogEvents", "logs:FilterLogEvents"],
         "Resource": ["arn:aws:logs:*:${ACCOUNT_ID}:log-group:/aws/emr-serverless/*"]
       },
       {
         "Sid": "EMRServerlessS3LogsAccess",
         "Effect": "Allow",
         "Action": ["s3:GetObject", "s3:ListBucket"],
         "Resource": "*"
       }
     ]
   }
   EOF
   ```

1. 連接政策：

   ```
   aws iam put-role-policy \
     --role-name SparkTroubleshootingMCPRole \
     --policy-name EMRServerlessTroubleshootingAccess \
     --policy-document file://emr-serverless-policy.json
   ```

### 選用：加密 CloudWatch Logs 的 KMS 許可
<a name="spark-troubleshooting-agent-iam-kms"></a>

如果您的 CloudWatch Logs 使用客戶管理的 KMS 金鑰加密，請新增下列項目 (`<KEY_ID>`以您的 KMS 金鑰 ID 取代）：

```
aws iam put-role-policy \
  --role-name SparkTroubleshootingMCPRole \
  --policy-name KMSCloudWatchLogsDecrypt \
  --policy-document "{
    \"Version\": \"2012-10-17\",
    \"Statement\": [{
      \"Effect\": \"Allow\",
      \"Action\": [\"kms:Decrypt\", \"kms:DescribeKey\"],
      \"Resource\": \"arn:aws:kms:${REGION}:${ACCOUNT_ID}:key/<KEY_ID>\"
    }]
  }"
```

## 步驟 3：設定您的 MCP 用戶端
<a name="spark-troubleshooting-agent-iam-step3"></a>

將 MCP 用戶端 （例如，Claude Desktop 或 Amazon Q Developer) 設定為使用您建立的角色 ARN：

```
echo "arn:aws:iam::${ACCOUNT_ID}:role/SparkTroubleshootingMCPRole"
```

請參閱 MCP 用戶端的文件，了解如何設定 AWS 登入資料 （通常是透過擔任此角色的 AWS 設定檔）。

## MCP 伺服器請求的條件索引鍵
<a name="spark-troubleshooting-agent-mcp-permissions-change"></a>

兩個條件索引鍵會自動新增至透過 SMUS MCP 伺服器提出的所有請求：
+ `aws:ViaAWSMCPService` – `true` 針對透過 AWS 受管 MCP 伺服器提出的任何請求，將 設定為 。
+ `aws:CalledViaAWSMCP` – 設定為 MCP 伺服器服務主體 （例如 `sagemaker-unified-studio-mcp.amazonaws.com`)。

當請求來自 AWS 受管 MCP 伺服器時，您可以使用這些條件金鑰來控制對 資源的存取。

**範例：僅允許透過 SMUS MCP 伺服器存取 Glue 讀取操作：**

```
{
  "Version": "2012-10-17",		 	 	 
  "Statement": [
    {
      "Sid": "AllowGlueReadViaSMUSMCP",
      "Effect": "Allow",
      "Action": ["glue:GetJob", "glue:GetJobRun", "glue:GetJobRuns"],
      "Resource": "*",
      "Condition": {
        "StringEquals": {
          "aws:CalledViaAWSMCP": "sagemaker-unified-studio-mcp.amazonaws.com"
        }
      }
    }
  ]
}
```

**範例：透過任何 AWS 受管 MCP 伺服器存取時拒絕刪除操作：**

```
{
  "Effect": "Deny",
  "Action": ["s3:DeleteObject", "s3:DeleteBucket"],
  "Resource": "*",
  "Condition": {
    "Bool": {
      "aws:ViaAWSMCPService": "true"
    }
  }
}
```

如需條件索引鍵的詳細資訊，請參閱《*IAM 使用者指南*》中的[AWS 全域條件內容索引鍵](https://docs.aws.amazon.com/IAM/latest/UserGuide/reference_policies_condition-keys.html)。