Operations Intelligence Platform

ADOstack

AI-Driven Operations

A suite of production-grade AI tools built for NOC, DevOps, and SRE teams.
Real clusters. Real incidents. Real intelligence.

6 Live Tools

6 Live Domains

View the Stack ↓

scroll

// The Stack

On-Call Assistant→

Incident Response & Escalation

The ADOStack operations hub. Aggregates live context from Infra Monitor, Incident Logger, and RAG Runbook in parallel when an alert fires, then generates a structured response plan. Manages full incident lifecycle with auto-resolution when metrics normalize, Telegram action buttons for high/critical, and shift handoff notes.

Claude API Python SQLite Flask Webhook Telegram HMAC Auth

Incident Orchestrator→

Multi-Agent Pipeline

A five-stage multi-agent pipeline that processes incidents end-to-end. Triage, Analyst, Escalation Gate, Runbook, and Summary agents collaborate via Claude — auto-remediating low-severity incidents and escalating critical ones to human review. Streams results live with Telegram dispatch on completion.

Claude API Python Multi-Agent Flask SSE Streaming RAG-Integrated Telegram

Incident Logger→

Detection & Record Keeping

Threshold detection engine for the ADOStack platform. Monitors CPU, memory, disk, and processes every 5 minutes — generates AI incident summaries via Claude and routes alerts to the On-Call Assistant via HMAC-signed webhook. Falls back to direct Telegram if On-Call is unavailable.

Claude API Python Flask SQLite Webhook HMAC Auth

RAG Runbook Assistant→

Retrieval-Augmented Generation

Ask natural-language questions against a corpus of ops runbooks. Semantic search via ChromaDB and MiniLM retrieves the most relevant procedures, then Claude generates grounded, cited answers.

Claude API Python ChromaDB MiniLM Flask RAG

Infra Monitor→

Infrastructure Monitoring

Real-time server health dashboard monitoring CPU, memory, disk, and network across VPS infrastructure. Claude generates plain-English NOC digests and anomaly alerts on demand.

Claude API Python Flask psutil Nginx systemd

K8s Event Summarizer→

Kubernetes Observability

Live Kubernetes cluster dashboard connected to a Civo k3s cluster. Tracks pod health, node status, deployments, and warning events. Claude synthesizes cluster state into actionable NOC digests.

Claude API Python Kubernetes Civo k3s Flask Nginx

// About

// What is ADOstack

ADOstack is a portfolio of production-deployed AI operations tools built to demonstrate real-world LLM integration across NOC, SRE, DevOps, and AI engineering workflows.

Every tool runs on live infrastructure — real clusters, real incident data, real API calls. No mockups. No local-only demos. Each project is deployed with Nginx, Gunicorn, systemd, and Cloudflare, mirroring the patterns used in production environments.

The stack grows with each project, covering RAG pipelines, Kubernetes observability, multi-agent orchestration, and LLM evaluation frameworks.

// Connect

⌥ github.com/ohdasdiego ◈ linkedin.com/in/diego-perez-51598967

// Stack Overview

Claude API Python Flask Kubernetes ChromaDB RAG Nginx Gunicorn systemd Cloudflare Ubuntu Civo