ongridio/ongrid
Ongrid is an AI-powered ops agent that autonomously investigates alerts, performs root-cause analysis, and orchestrates fixes for IT infrastructure issues, interacting via chat platforms.
Awesome AI for Infra › Incident Management & Response
papadopouloskyriakos/agentic-chatops
Agentic-ChatOps is an AI-driven ChatOps system designed to autonomously manage complex IT infrastructure. It features three agentic subsystems: ChatOps (infrastructure), ChatSecOps (security), and ChatDevOps (CI/CD), built on n8n orchestration and Matrix for human interaction. The core innovation lies in its tiered agent architecture, which progresses from deterministic triage scripts to sophisticated AI models like Claude Code, always keeping a human in the loop for approval before any action is taken. The project introduces several advanced AI capabilities, including a 'Self-Improving Prompts' mechanism that uses an LLM-as-a-Judge to evaluate performance and dynamically patch prompts through A/B trials. At its heart is the 'Infragraph', a causal world model of the infrastructure with a non-bypassable prediction gate; this ensures that all remediation proposals are accompanied by machine-computed consequence predictions based on a detailed dependency graph. This system combines AI planning with proven Ansible playbooks, allowing AI agents to generate investigation plans and execute reliable automation. It also incorporates predictive alerting based on trending risk across devices, 5-signal RAG (Retrieval Augmented Generation) with GraphRAG for enhanced knowledge retrieval, and a Karpathy-style compiled knowledge base for comprehensive context. The system is equipped with a full observability stack using OpenTelemetry and a robust formal evaluation pipeline to continuously assess and improve its performance.
https://github.com/papadopouloskyriakos/agentic-chatops
Also in Incident Management & Response
Tommy-yw/RunbookHermes
RunbookHermes is an AIOps agent built on the Hermes Agent framework, designed for evidence-driven incident response, approval-gated remediation, and self-evolving runbook learning.
Arvo-AI/aurora
Aurora is an open-source, AI-powered incident management platform that uses AI agents to autonomously investigate incidents, perform root cause analysis, and suggest remediations across multi-cloud...
swapnildahiphale/OpenSRE
OpenSRE is an open-source AI SRE agent that automates incident investigation, root cause analysis, and learns from past incidents using episodic memory and a knowledge graph.
olemeyer/rocketplaneIO
AI SRE for Kubernetes providing zero-instrumentation eBPF observability, an AI copilot for incident remediation via guardrailed, self-verifying actions, and air-gapped capable operations.