How hik8s/core works
Project type and purpose
• This is a microservices‐style data‐processing and backend system for a Kubernetes chatbot (“HiK8s Core”), implemented in Rust.
• It consists of multiple standalone services (each with its own Cargo.toml and main.rs), plus a shared library and end-to-end tests.
File structure
• Top‐level “rs/” directory holds subprojects:
• data-intake (HTTP ingest service)
• data-processing (event processing workers)
• data-vectorizer (embeddings/vectorization)
• chat-backend (chat API server)
• shared (common code: DB connectors, message-broker clients, OpenAI client, routing, types)
• tests (E2E integration tests)
• CI/CD workflows and Dockerfile at root.
• Config files for Rocket (chat-backend), Clippy, Buck2, plus Docker and README.
README for additional insights
• External dependencies: GreptimeDB (time-series), Qdrant (vector DB), Redis (cache), Fluvio (message broker), OpenAI.
• Data flows: ingest → processing → vectorize → store/query → chat API.
Detail
• Microservice as its own component
• Intermediate Fluvio topics if desired (e.g., “raw-logs”, “processed-events”)
• Authentication/authorization if part of shared router (e.g., JWT guard)
• Retry or error-handling paths if notable