将 CDK 与 AppSync 结合使用的 DynamoDB 事务

Question

我正在尝试使用 DynamoDB 作为需要在其他 DynamoDB 表上执行事务的数据源来解决突变

mutation MyMutation {
  like(tweetId: "0ABCDEFGHIK")
}

错误：

{
  "data": null,
  "errors": [
    {
      "path": [
        "like"
      ],
      "data": null,
      "errorType": null,
      "errorInfo": null,
      "locations": [
        {
          "line": 2,
          "column": 3,
          "sourceName": null
        }
      ],
      "message": "Failed to execute DynamoDB transaction"
    }
  ]
}

即当我的喜欢表上发生突变时，我需要在我的用户和推文表上执行写入和更新。

根据说明，appync 需要承担执行交易的角色，我已经实现了以下结构

服务角色

const likeTableTransactionItemRole = new iam.Role(
  this,
  "likeTableTransactionItemRole",
  {
    assumedBy: new iam.ServicePrincipal("appsync.amazonaws.com"),
  }
);

附加内联策略（Ps：我知道 * 是不好的做法，目标是让它发挥作用）

likeTableTransactionItemRole.attachInlinePolicy(
  new iam.Policy(this, "likeTableTransactionItemPolicy", {
    statements: [
      new iam.PolicyStatement({
        effect: iam.Effect.ALLOW,
        actions: ["*"],
        resources: [likesTable.tableArn, likesTable.tableArn + "/*"],
      }),
      new iam.PolicyStatement({
        effect: iam.Effect.ALLOW,
        actions: ["*"],
        resources: [
          tweetsTable.tableArn,
          tweetsTable.tableArn + "/*",
          usersTable.tableArn,
          usersTable.tableArn + "/*",
        ],
      }),
    ],
  })
);

为其创建 DynamDB 数据源

const transactionDataSource = new DynamoDbDataSource(
  this,
  "likesTableTransaction",
  {
    api,
    table: likesTable,
    serviceRole: likeTableTransactionItemRole,
  }
);

为突变创建解析器

 const likeMutationResolver = new appsync.Resolver(this, "likes", {
      api,
      typeName: "Mutation",
      fieldName: "like",
      dataSource: transactionDataSource,
      requestMappingTemplate: appsync.MappingTemplate.fromFile(
        path.join(mappingTemplatesLocation, "Mutation.like.request.vtl")
      ),
      responseMappingTemplate: appsync.MappingTemplate.fromFile(
        path.join(mappingTemplatesLocation, "Mutation.like.response.vtl")
      ),
    });

任何线索将不胜感激