papadopouloskyriakos

papadopouloskyriakos / agentic-chatops

Public

3-tier agentic ChatOps (n8n + GPT-4o + Claude Code) implementing all 21 patterns from "Agentic Design Patterns" — solo operator managing 137 devices

kyriakos.papadopoulos.tech agentic-ai chatops claude-code devops infrastructure-automation
10
1
69% credibility
Found Mar 24, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Shell
AI Summary

Self-hosted AI agents automatically triage infrastructure alerts in chat rooms, investigate issues, propose fixes for human approval, and execute them safely while keeping people in the loop.

How It Works

1
😴 Waking up to alerts?

You're the solo operator handling hundreds of servers and alerts at 3am across multiple sites.

2
⚙️ Add your AI helper

Connect it once to your chat app and monitoring so it watches everything for you.

3
🚨 Alert hits the chat

A problem like a downed server pings your chat room automatically.

4
🔍 AI digs in right away

Your smart sidekick checks status, logs, and past fixes without bugging you.

5
📋 Findings and plan appear

AI shares a clear summary of what went wrong and a safe step-by-step fix.

6
👍 You give quick approval

React with thumbs up, pick from a poll, or reply – always your call.

✅ Fixed and done

AI handles the repair, verifies it worked, clears the alert – sleep easy.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 10 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is agentic-chatops?

This Shell-based ChatOps platform automates infrastructure triage for a solo operator handling 137 devices across 6 sites. Alerts from LibreNMS or Prometheus hit a 3-tier agentic system: GPT-4o handles fast L1 triage (dedup, investigate, score confidence), Claude Code dives into L2 analysis with remediation plans via Matrix polls, and humans approve Tier 3 execution. It implements all 21 patterns from "Agentic Design Patterns" like reflection, tool use, and human-in-the-loop, turning 3am alerts into structured YouTrack issues with AI proposals.

Why is it gaining traction?

Unlike basic alert bots, this 3-tier architecture delivers end-to-end automation—80% alerts resolved without escalation—while enforcing budgets ($5/session) and guardrails like safe exec wrappers. Developers dig the real-world proof: one person manages 310 objects (VMs, K8s, switches) using n8n workflows, MCP tools, and Matrix commands like !issue start or !mode oc-cc. It's a battle-tested 3-tier project GitHub example blending GPT-4o speed with Claude reasoning.

Who should use this?

Solo DevOps engineers or homelabbers drowning in alerts from Proxmox clusters, K8s nodes, or multi-site networks. Ideal for those running self-hosted stacks (Matrix, GitLab, Grafana) without a team, needing AI to triage bursts across 137 devices before human polls.

Verdict

Grab it for inspiration on agentic ChatOps—docs, metrics, and E2E tests impress despite 10 stars and 0.699999988079071% credibility score signaling early maturity. Fork and adapt if you're building 3-tier agentic flows; production-ready for solo ops but tweak for scale.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.