
Pinecone Assistant MCP Server を使ってみた
こんにちは!クラウド事業本部コンサルティング部のたかくに(@takakuni_)です。
少し前に、 Pinecone Database についての MCP Server を触ってみたブログを書きました。
どうやら、Pinecone Assistant 側にも、 MCP Server があるようで、今回はこちらに試してみたいと思います。
Pinecone Assistant MCP Server
Pinecone Assistant MCP Server は Pinecone Assistant と MCP を通じて、会話する MCP Server です。主な機能としては、Pinecone Assistant に蓄積されたナレッジから、自然言語でクエリを実行します。
起動方法
ドキュメントを辿ると、Remote MCP, Local MCP の 2 種類をサポートしていました。
Remote MCP では、Pinecone Assistant に割り当てられた MCP Endpoint 宛てに直接リクエストを送信し、Local MCP では Docker でホストした MCP サーバー経由で Pinecone Assistant にアクセスします。
Remote MCP(LangChain MCP client を利用したケース)
# Example code for integrating with LangChain
from langchain_mcp_adapters.client import MultiServerMCPClient
from langgraph.prebuilt import create_react_agent
from langchain_anthropic import ChatAnthropic
model = ChatAnthropic(model_name="claude-3-7-sonnet-latest", api_key=<YOUR_ANTHROPIC_API_KEY_HERE>)
pinecone_api_key = "<YOUR_PINECONE_API_KEY>"
async with MultiServerMCPClient(
{
"assistant_ai_news": {
# Pinecone Assistant に割り当てられた MCP Endpoint を指定
"url": "https://prod-1-data.ke.pinecone.io/mcp/assistants/ai-news/sse",
"transport": "sse", # SSE を利用
"headers": {
"Authorization": f"Bearer {pinecone_api_key}"
}
},
"assistant_industry_reports": {
"url": "https://prod-1-data.ke.pinecone.io/mcp/assistants/industry-reports/sse",
"transport": "sse",
"headers": {
"Authorization": f"Bearer {pinecone_api_key}"
}
}
}
) as client:
agent = create_react_agent(model, client.get_tools())
response = await agent.ainvoke({"messages": "Your task is research the next trends in AI, and form a report with the most undervalued companies in the space.
You have access to two assistants, one that can help you find the latest trends in AI, and one that can help you find reports on companies."})
print(response["messages"][-1].content)
Local MCP(Docker から起動)
Remote MCP では、Pinecone Assistant に割り当てられた MCP Endpoint を指定しますが、PINECONE_ASSISTANT_HOST には Pinecone Assistant のホスト名を入力します。
{
"mcpServers": {
"pinecone-assistant": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"PINECONE_API_KEY",
"-e",
"PINECONE_ASSISTANT_HOST",
"pinecone/assistant-mcp"
],
"env": {
"PINECONE_API_KEY": "<YOUR_PINECONE_API_KEY>",
"PINECONE_ASSISTANT_HOST": "<YOUR_PINECONE_ASSISTANT_HOST>"
}
}
}
}
やってみる
早速、Pinecone Assistant MCP Server を利用してみたいと思います。
Assistant の作成
まずは、Assistant の作成を行います。
名前は mcp-assistant
としました。
後ほど利用するホスト名をメモしておきます。
ドキュメントの登録
ドキュメントを登録します。ドラッグ&ドロップで PDF ファイルを登録します。
今回は以下のドキュメントを登録しました。
試しに Web 上で質問をしてみましたが、うまく回答できているようです。
コンテナイメージのビルド
ローカルで MCP サーバーを起動するためにコンテナイメージをビルドします。
git clone https://github.com/pinecone-io/assistant-mcp.git
cd assistant-mcp
docker build -t pinecone/assistant-mcp .
コンテナイメージが作れていることを確認します。
takakuni@ assistant-mcp % docker image ls -a
REPOSITORY TAG IMAGE ID CREATED SIZE
pinecone/assistant-mcp latest 26ea6e44e488 7 seconds ago 114MB
MCP サーバーの立ち上げ
わたしは Roo Code をよく利用しているため、 mcp.json に以下を追記します。PINECONE_API_KEY, PINECONE_ASSISTANT_HOST は適宜、設定しましょう。
{
"mcpServers": {
"pinecone-assistant": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"PINECONE_API_KEY",
"-e",
"PINECONE_ASSISTANT_HOST",
"pinecone/assistant-mcp"
],
"env": {
"PINECONE_API_KEY": "<YOUR_PINECONE_API_KEY_HERE>",
"PINECONE_ASSISTANT_HOST": "<YOUR_PINECONE_ASSISTANT_HOST_HERE>"
}
}
}
}
設定完了すると MCP サーバーがオンになっていますね。
質問
先ほどと同じかつ、MCP Server の挙動を確認したいため、「AWS Fargate Spot を利用するメリットを教えてください Pinecone MCP Server から取得してください。」と質問しました。
意図した通り、 MCP サーバーの呼び出しが発生していますね。
ただ、Assistant Name のパラメーターに誤りがあり、リクエストに失敗している様子でした。
Pinecone Assistant MCP Server を利用する場合は、 Pinecone Assistant の名前を意識して作る必要があるようです。
Assistant Name を教えて、実行してみました。
マスクしていますが、 Pinecone MCP Server とは違い、検索結果に signed_url
(署名付き URL)が発行されていました。ソースドキュメントに設定できそうな予感を感じます。
Pinecone Assistant MCP Server の検索結果
{"content":"© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate ⼊⾨吉⽥ 英史Solutions Architect2024/10\n1\nAWS Black BeltOnline Seminar© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\n⾃⼰紹介吉⽥英史アマゾンウェブサービスジャパンソリューションアーキテクト⼩売・消費財のお客様を中⼼にご⽀援しています。⽣活に⽋かせない様々なビジネスをクラウドで加速するお⼿伝いができることを、何より嬉しく感じています。好きなAWSサービスAWS Fargate\n 2© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\n本セミナーの主な対象者•これからAWSを利⽤される予定の、アプリケーションおよびインフラ担当者•サーバレスコンテナ実⾏環境である、AWS Fargate の概要や始め⽅を知りたい⽅•クラウド上の既存ワークロードの、コンテナ化を検討している⽅•オンプレミスの既存コンテナワークロードの、クラウド移⾏を検討している⽅\n3© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\n1.AWS Fargate とは2.AWS Fargate とAmazon EC2の違い3.AWS Fargate の始め⽅4.まとめ\nアジェンダ\n4© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate とは\n5© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS のコンテナサービス\nAWSCloud Map\nオーケストレーションコンテナのデプロイ、スケジューリング、スケーリングイメージレジストリコンテナイメージの格納\nホスティングコンテナ実⾏環境その他の関連サービス\nAmazon CloudWatch Container Insights\nAmazon ECS Amazon EKS Amazon ECR\nAmazon EC2 AWS Fargate\nAWS App Runner\n6© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nコンテナ実⾏環境の選択肢\n7\nAmazon ECS\nAmazon EC2AWS Fargate\nコンテナオーケストレーター(コントロールプレーン)\nコンテナの実⾏環境(データプレーン)\nAmazon EKS\n© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate はサーバレスのコンテナ実⾏環境\n8\n•ホストのアップグレード、パッチ適⽤が対応不要•クラスターのキャパシティ管理が不要•設計による分離とセキュリティの担保•前払い料⾦なし、利⽤リソースに応じた従量課⾦•Fargate SpotとFargate Gravitonによるコスト削減※ Amazon EKS では未サポート\n※ Compute Savings Plans 対象© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS FargateSpot(ECS のみ)•AWS Fargate で中断処理に強いワークロードを実⾏するためのオプション•通常のFargateの価格と⽐較して最⼤ 70% 割引•Capacity Provider のFARGATE_SPOT キャパシティとして利⽤可能•AWS Graviton も利⽤可能\n9\nECS Cluster\nAWS FargateAWS FargateSpot\nCapacityProviderFARGATECapacityProviderFARGATE_SPOT\nhttps://docs.aws.amazon.com/ja_jp/AmazonECS/latest/developerguide/fargate-capacity-providers.htmlhttps://aws.amazon.com/jp/about-aws/whats-new/2024/09/amazon-ecs-graviton-based-spot-compute-fargate/© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nECS/EKS におけるコンテナ実⾏環境の⽐較\n10\nEKS on EC2(セルフマネージド型ノード)\nECS on EC2\nEKS on EC2(マネージド型ノードグループ)\nECS on Fargate\nEKS on Fargate\nホストのスケーリング\nホストの OS / ライブラリ設定コンテナエージェント設定\nアプリケーションコンテナホストのスケーリング\nホストの OS / ライブラリ設定コンテナエージェント設定\nアプリケーションコンテナホストのスケーリング\nホストの OS / ライブラリ設定コンテナエージェント設定\nアプリケーションコンテナ\nAWSが提供するレイヤーお客様に管理いただくレイヤー\n抽象化のレベル\nFargateを活⽤することで、アプリケーションコンテナの開発に集中できる© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAmazon ECS/EKS on Fargate の動作イメージ\nAWS Fargate\nスケジューリングとオーケストレーションクラスター管理配置エンジン\nAmazon ECS/EKS\nAWS Fargate\nタスク/Pod\nクラスター\nAWS Fargate\n11\nタスク/Podタスク/Pod\nタスク/Podタスク/Podタスク/Pod\n© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate とAmazon EC2の違い\n12© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate とAmazon EC2で異なる 5 つの観点1.vCPU とメモリ2.ホスト3.ネットワークモード4.データボリューム5.セキュリティ\nAmazon EC2との違い\n13\nAmazon EC2AWS Fargate︖© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate とAmazon EC2で異なる 5 つの観点1.vCPU とメモリ2.ホスト3.ネットワークモード4.データボリューム5.セキュリティ\nAmazon EC2との違い\n14\nAmazon EC2AWS Fargate︖© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\n柔軟な設定の選択肢 : 74 パターンのvCPU, メモリの組み合わせから選択可能\nタスク/Pod に割り当てるvCPU とメモリ設定vCPU Memory256 (.25 vCPU)0.5GB, 1GB, 2GB →3 種類512 (.5 vCPU)1GB ~ 4GB (1GB 刻み) →4 種類1,024 (1 vCPU)2GB ~ 8GB (1GB 刻み) →7 種類2,048 (2 vCPU)4GB ~ 16GB (1GB 刻み) →13 種類4,096 (4 vCPU)8GB ~ 30GB (1GB 刻み) →23 種類8,192 (8 vCPU)16GB ~ 60GB (4GB 刻み) →12 種類16,384 (16 vCPU)32GB ~ 120GB (8GB 刻み) →12 種類© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate とAmazon EC2で異なる 5 つの観点1.vCPU とメモリ2.ホスト3.ネットワークモード4.データボリューム5.セキュリティ\nAmazon EC2との違い\n16\nAmazon EC2AWS Fargate︖© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nホスト環境の違いによるFargate の制約 1/2 •GPU など、ハードウェアアクセラレーターが未サポート (2024年9⽉時点)•デーモンスケジューリング戦略 (ECS)、デーモンセット (EKS) はサポート外→サイドカーパターンに\n17\nWeb ServerLog Router\nEC2 Instance\nWeb ServerLog Router\nLog Router\nTask/PodTask/PodFargate\nデーモン サイドカー\nWeb Server\nWeb Server\nTask/PodTask/Pod\nTask/Pod© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nホスト環境の違いによるFargate の制約 2/2 •Fargate はホストを","reference":{"file":{"created_on":"2025-05-08T07:49:39.310977396Z","error_message":null,"id":"11f4bb3e-b6af-45d5-9711-995f3f483034","metadata":null,"name":"AWS-Black-Belt_2024_AWS-Fargate-Introduction_1031_v1.pdf","percent_done":1.0,"signed_url":"masked","size":3021798,"status":"Available","updated_on":"2025-05-08T07:49:53.155676273Z"},"pages":[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18],"type":"pdf"},"score":0.99773055,"type":"text"}
{"content":"るhttps://aws.amazon.com/jp/blogs/news/reducing-aws-fargate-startup-times-with-zstd-compressed-container-images/AWS Fargate はシーク可能なOCI を使⽤してより⾼速なコンテナ起動を可能にhttps://aws.amazon.com/jp/blogs/news/aws-fargate-enables-faster-container-startup-using-seekable-oci/\nEC2 Instance\nキャッシュ\nコンテナ\nイメージレジストリ\nFargate\n2度⽬のコンテナ起動時© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate とAmazon EC2で異なる 5 つの観点1.vCPU とメモリ2.ホスト3.ネットワークモード4.データボリューム5.セキュリティ\nAmazon EC2との違い\n19\nAmazon EC2AWS Fargate︖© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\n利⽤可能なネットワークモード\n20\nFargate実⾏環境\nhttps://docs.aws.amazon.com/ja_jp/AmazonECS/latest/bestpracticesguide/networking-networkmode.html\nEC2実⾏環境\nawsvpc モードoECS/EKS 管理下のENI(仮想サーバーの仮想ネットワークカード)がタスクにアタッチされる\nbridge モードo仮想ネットワークブリッジを利⽤してホスト/コンテナポートをマッピングして通信\nhost モードoコンテナをホストしているEC2 インスタンスのIP/ポートを介して通信EC2 instance\nENI\nPort 3000IP x.x.x.xPort 3000\nEC2 instance\nENI\nPort 3000\nPort 3000\nIP x.x.x.xPort 50077\nIP x.x.x.xPort 52330\nEC2 instance\nENI\nPort 3000\nPort 3000\nIP x.x.x.xPort 3000\nIP y.y.y.yPort 3000\nENI© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate とAmazon EC2で異なる 5 つの観点1.vCPU とメモリ2.ホスト3.ネットワークモード4.データボリューム5.セキュリティ\nAmazon EC2との違い\n21\nAmazon EC2AWS Fargate︖© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\n利⽤可能なデータボリューム\n22\nFargate バインドマウント/ エフェメラルストレージo揮発性のストレージをコンテナにマウントし、コンテナ間で共有o使⽤するすべてのコンテナが停⽌するとデータが削除されるo20GiBまで無料、追加分にのみ課⾦ (ECS Max 200GiB / EKS Max 175GiB)\nAmazon EFS ボリュームoストレージ容量が伸縮⾃在で、⾃動的に拡⼤および縮⼩される\nFargate実⾏環境Amazon EBS ボリューム (Fargate はECSのみ)o⾼スループットなトランザクション集約型アプリケーション向けoタスクごとに1つの EBS ボリュームにアタッチ可能\nAmazon ECS: https://docs.aws.amazon.com/ja_jp/AmazonECS/latest/developerguide/using_data_volumes.htmlAmazon EKS: https://docs.aws.amazon.com/ja_jp/eks/latest/userguide/storage.html© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate とAmazon EC2で異なる 5 つの観点1.vCPU とメモリ2.ホスト3.ネットワークモード4.データボリューム5.セキュリティ\nAmazon EC2との違い\n23\nAmazon EC2AWS Fargate︖© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nセキュリティにおけるFargate のメリット•OS、コンテナエンジン、エージェントなどのパッチ当てが不要•タスク/Podごとに実⾏環境が分離されている•コンテナに特権モードが使⽤できない•awsvpc ネットワークモードのためENI や SG がタスク/Pod ごとに独⽴•sshなどによるホストへのアクセスが不可能\n24\nFargate セキュリティのベストプラクティスhttps://docs.aws.amazon.com/ja_jp/AmazonECS/latest/developerguide/security-fargate.html\nFargateKernel\nVMTask/Pod\nKernel\nVMTask/Pod\nEC2Kernel\nVMTask/PodTask/PodTask/Pod\n© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nFargate 上のコンテナのデバッグ (ECS のみ)•Fargate 上のコンテナにログインするにはECS Exec を使⽤•AWS Systems Manager セッションマネージャーを使⽤するため、事前にプラグインの導⼊が必要\n25\n※ 2024/09 時点でコンソールからの有効化は未対応ECS Exec を使⽤してAmazon ECS コンテナをモニタリングするhttps://docs.aws.amazon.com/ja_jp/AmazonECS/latest/developerguide/ecs-exec.html\nCLI,SDKなど\naws ecs execute-command ¥ --cluster cluster-name ¥ --task task-id ¥ --container container-name ¥ --interactive ¥ --command \"/bin/sh\"\nServiceAWS Fargate Task\n有効化\nECS Exec© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate の始め⽅\n26© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate の始め⽅•AWS Fargate の開始⽅法–https://aws.amazon.com/jp/fargate/getting-started/•AWS Fargate の料⾦–https://aws.amazon.com/jp/fargate/pricing/•Amazon ECS/AWS Fargate 利⽤構成と料⾦試算例–https://aws.amazon.com/jp/cdp/ec-container/•AWS Fargate のお客様導⼊事例–https://aws.amazon.com/jp/containers/customers/•AWS Fargate に関するAWS Blog 記事–https://aws.amazon.com/jp/blogs/news/tag/aws-fargate/ 27\n© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nまとめ•AWS Fargateとは§サーバレスのコンテナ実⾏環境§コンテナクラスターの運⽤負荷を軽減することが可能•AWS Fargate の、Amazon EC2と異なる5 つの観点§vCPUとメモリ / ホスト /ネットワーク /ボリューム / セキュリティ§タスク定義の詳細•Amazon ECS の始め⽅\n28\n© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Black Belt Online Seminar とは•「サービス別」「ソリューション別」「業種別」などのテーマに分け、アマゾン ウェブ サービス ジャパン合同会社が提供するオンラインセミナーシリーズです•AWS の技術担当者が、 AWS の各サービスやソリューションについてテーマごとに動画を公開します•以下の URL より、過去のセミナー含めた資料などをダウンロードすることができます–https://aws.amazon.com/jp/aws-jp-introduction/aws-jp-webinar-service-cut/–https://www.youtube.com/playlist?list=PLzWGOASvSx6FIwIC2X1nObr1KcMCBBlqY\n29\nご感想は X (Twitter) へ︕ハッシュタグは以下をご利⽤ください#awsblackbelt \n© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\n内容についての注意点•本資料では資料作成時点のサービス内容および価格についてご説明しています。AWS のサービスは常にアップデートを続けているため、最新の情報は AWS 公式ウェブサイト (https://aws.amazon.com/) にてご確認ください•資料作成には","reference":{"file":{"created_on":"2025-05-08T07:49:39.310977396Z","error_message":null,"id":"11f4bb3e-b6af-45d5-9711-995f3f483034","metadata":null,"name":"AWS-Black-Belt_2024_AWS-Fargate-Introduction_1031_v1.pdf","percent_done":1.0,"signed_url":"masked","size":3021798,"status":"Available","updated_on":"2025-05-08T07:49:53.155676273Z"},"pages":[18,19,20,21,22,23,24,25,26,27,28,29,30],"type":"pdf"},"score":0.997459,"type":"text"}
{"content":"ask/PodTask/PodFargate\nデーモン サイドカー\nWeb Server\nWeb Server\nTask/PodTask/Pod\nTask/Pod© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nホスト環境の違いによるFargate の制約 2/2 •Fargate はホストを都度作り直すため、コンテナイメージを Pull する際にローカルキャッシュが使⽤できない→コンテナ起動時間の短縮のアイデア•イメージサイズの最適化•イメージの圧縮•遅延読み込みSeekable OCI (SOCI) (ECS のみ)\n18\nキャッシュ\nコンテナ\nzstd 圧縮したコンテナイメージを使⽤してAWS Fargate の起動時間を短縮するhttps://aws.amazon.com/jp/blogs/news/reducing-aws-fargate-startup-times-with-zstd-compressed-container-images/AWS Fargate はシーク可能なOCI を使⽤してより⾼速なコンテナ起動を可能にhttps://aws.amazon.com/jp/blogs/news/aws-fargate-enables-faster-container-startup-using-seekable-oci/\nEC2 Instance\nキャッシュ\nコンテナ\nイメージレジストリ\nFargate\n2度⽬のコンテナ起動時© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nAWS Fargate とAmazon EC","reference":{"file":{"created_on":"2025-05-08T07:49:39.310977396Z","error_message":null,"id":"11f4bb3e-b6af-45d5-9711-995f3f483034","metadata":null,"name":"AWS-Black-Belt_2024_AWS-Fargate-Introduction_1031_v1.pdf","percent_done":1.0,"signed_url":"masked","size":3021798,"status":"Available","updated_on":"2025-05-08T07:49:53.155676273Z"},"pages":[17,18,19],"type":"pdf"},"score":0.80730873,"type":"text"}
{"content":"https://aws.amazon.com/) にてご確認ください•資料作成には⼗分注意しておりますが、資料内の価格と AWS 公式ウェブサイト記載の価格に相違があった場合、AWS 公式ウェブサイトの価格を優先とさせていただきます•価格は税抜表記となっています。⽇本居住者のお客様には別途消費税をご請求させていただきます•技術的な内容に関しましては、有料のAWSサポート窓⼝へお問い合わせください•料⾦⾯でのお問い合わせに関しましては、カスタマーサポート窓⼝へお問い合わせください (マネジメントコンソールへのログインが必要です) \n30© 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.\nThank you!\n31","reference":{"file":{"created_on":"2025-05-08T07:49:39.310977396Z","error_message":null,"id":"11f4bb3e-b6af-45d5-9711-995f3f483034","metadata":null,"name":"AWS-Black-Belt_2024_AWS-Fargate-Introduction_1031_v1.pdf","percent_done":1.0,"signed_url":"masked","size":3021798,"status":"Available","updated_on":"2025-05-08T07:49:53.155676273Z"},"pages":[30,31],"type":"pdf"},"score":0.08195955,"type":"text"}
Pinecone MCP Server の検索結果
まとめ
以上、「Pinecone Assistant MCP Server を使ってみた」でした。
インデックス or Assitant といった形でデータソースが異なりますが、やっていることは、ほとんど同じに思えました。
シンプルな RAG の検索のみであれば、Pinecone MCP Server、署名付き URL など参考情報を見せたいケースでは、Pinecone Assistant MCP Server の使い分けが良さそうに思えます。
クラウド事業本部コンサルティング部のたかくに(@takakuni_)でした!