Using a Knowledge Graph to implement a DevOps RAG application

RAG (Retrieve, Answer, Generate) applications are becoming popular for creating company documentation chatbots and similar tools. Typically, these applications rely on unstructured text as their source of knowledge. However, not all information arrives in this format. In the case of a microservice architecture, tasks are mostly defined as unstructured text. To prepare information about such an architecture, one option is to create daily snapshots and transform them into text that a Language Learning Model (LLM) can understand. Another approach is using knowledge graphs, which can store both structured and unstructured information in a single database. Nodes and relationships are used to describe data in a knowledge graph, with nodes representing entities or concepts like people, organizations, locations, etc., and relationships defining connections between these entities. Knowledge graphs allow users to store and retrieve both structured and unstructured information to power their RAG applications. In this blog post, the author demonstrates how to implement a knowledge graph-based RAG application with LangChain to support a DevOps team. The code is available on GitHub.