
Show HN: AIOps Cockpit – An open-source AI SRE with human-in-the-loop safety
The open-source, AI-powered SRE control room.
以下正文同步自 Hacker News 最新,版权归原站所有,已转换为易读排版。
The open-source, AI-powered SRE control room.
Diagnose incidents in seconds with an AI agent that reads your logs, checks your metrics, and proposes a fix. 3D visualization. Human-in-the-loop safety. Self-hosted.
3D Spatial Topology Map - See your infrastructure as a living, breathing 3D space. Green is healthy, red is on fire, spinning means failing.
AI Agent Diagnostics - Describe the incident ("checkout API 500s, DB restarting") and the LLM (Claude 3.7) investigates using real tools.
Human-in-the-Loop Safety - The AI never applies a fix directly. It generates a unified diff. A human must review and approve it. Requires a two-person rule (requester cannot self-approve).
Defense-in-Depth - Append-only audit logs, prompt-injection defenses, and a strict tool allowlist. Destructive actions are disabled by default.
Hand Gestures (Optional) - Wave your hand to control the camera via MediaPipe (pinch = zoom, wrist = rotate, point = select).
Real-time Telemetry - Live host metrics (CPU, RAM, Disk) and Docker container states via Socket.IO.
Distributed Telemetry: Install the lightweight remote agent on any server (AWS, GCP, homelab) and securely stream its Docker and host metrics back to the central 3D cockpit.
Quick Start
Prerequisites: Node.js 20+ and pnpm.
正文由 探How 从来源站点 RSS 同步,内容未经改写;遇到排版缺失或需要图片、视频时请以原文为准。