使用 Terraform 构建 Bedrock 知识库：在 AWS 上实现首个 RAG 管道
📄 中文摘要
Bedrock 知识库负责内容的分块、嵌入和检索，减轻了用户的负担。然而，底层基础设施如 OpenSearch Serverless、S3 和 IAM 策略需要通过 Terraform 来确保生产环境的准备。虽然用户可以通过 Bedrock 端点回答一般性问题，但在询问公司内部文档时，模型可能会产生不准确的回答。RAG（检索增强生成）技术能够将模型的响应与实际数据相结合，从而填补这一空白。AWS Bedrock 知识库是一项完全托管的 RAG 服务，用户只需将其指向包含文档的 S3 存储桶，系统便会自动处理内容分块、生成嵌入、在 OpenSearch Serverless 中存储向量，并在查询时进行检索，无需自定义嵌入管道、向量数据库管理或编写检索逻辑。
📄 English Summary
Bedrock Knowledge Base with Terraform: Your First RAG Pipeline on AWS 🔍
The Bedrock Knowledge Base manages chunking, embedding, and retrieval, relieving users of these tasks. However, the underlying infrastructure, including OpenSearch Serverless, S3, and IAM policies, requires Terraform to be production-ready. While users have a Bedrock endpoint that can answer general questions, it may hallucinate confidently when asked about internal company documents. Retrieval-Augmented Generation (RAG) fills this gap by grounding model responses in actual data. AWS Bedrock Knowledge Bases is a fully managed RAG service that automatically handles content chunking, embedding generation, vector storage in OpenSearch Serverless, and retrieval at query time when pointed at an S3 bucket containing documents, eliminating the need for custom embedding pipelines, vector database management, or retrieval logic.