How chriskalix/hades works
Project type and purpose
• Hades is a distributed Host‐based Intrusion Detection System (HIDS) that consists of:
• An Agent component (runs on each monitored host, uses eBPF and netlink to collect system events).
• A Plugin subsystem (drivers, collectors, network control, egress guard) for modular data acquisition and enforcement.
• A Server component that receives events (via gRPC/transfer), persists them (MongoDB, Redis), analyzes the data, and exposes a Web Console/UI.
File structure
• Top‐level dirs:
• SDK (Go & Rust libraries for clock, config, logger, transport, hashing, downloading).
• agent (Go agent code, plugins manager, metrics, transport, proto definitions, deploy scripts).
• plugins (collector in Go, edriver in Go, eguard in Rust, ncp in Go, libs).
• server (Webconsole API & static assets, gRPC handlers, transfer service, pkg for Mongo/Redis, config).
• Key patterns:
• “agent → transport → transfer → gRPC server → storage → webconsole.”
• Plugin‐based architecture (each plugin under plugins/ is an independent module).
• SDK abstracts cross‐cutting concerns (logging, transport, config).
README for additional insights
• Agent data‐flow: kernel probes → BPF programs → user‐space plugin hooks → transport client.
• Data Analysis: transfer pool → gRPC server → handlers → data persistence → UI/analytics.
Elements
• Clear labels for each component (e.g., “Agent Core,” “Collector Plugin,” “gRPC Transfer Service”)
Host → Transport → Transfer Server → gRPC → Storage → API → UI
• Important config & cert files near gRPC Transfer.