AutoOps AI

Configure observability and investigate incidents for any codebase with Splunk and AI.

AutoOps AI is an autonomous observability engineer. Install it, connect an LLM provider, point it at any repository, and it scans your architecture, generates instrumentation, bootstraps Splunk and OpenTelemetry in Docker, creates dashboards and alerts, and runs AI-driven incident investigations with root cause analysis and remediation recommendations.

Repository: https://github.com/kenzzhood/AutoOps
PyPI: https://pypi.org/project/autoops-ai/

---

Table of Contents

• Architecture
• Features
• Prerequisites & Dependencies
• Install
• Quick Start
• How It Works
• Command Reference
• Configuration
• Example Apps & Validation
• Development
• Built With
• License

---

Architecture

Required submission diagram: architecture_diagram.png
Detailed write-up: architecture_diagram.md

Complete Flow

---

Features

• Multi-provider LLM — OpenAI, Claude, Azure OpenAI, OpenRouter, Amazon Bedrock
• Configure pipeline — autoops configure --repo . bootstraps the full observability stack
• Splunk bootstrap — Auto-starts Splunk Enterprise in Docker (autoops-splunk)
• OpenTelemetry collector — Traces, metrics, and logs pipeline (autoops-otel-collector)
• Architecture discovery — LLM scans your repo and maps services, APIs, databases
• Auto-instrumentation — Generates autoops_middleware.py and autoops_db_tracing.py
• Splunk artifacts — Dashboards, saved searches, and alerts out of the box
• AI incident pipeline — Evidence collection → RCA → Remediation with incident reports

---

Prerequisites & Dependencies

System requirements

Requirement	Version	Purpose
Python	3.11+	CLI and agents
Docker Desktop	Latest	Splunk + OTel containers
LLM API key	Any supported provider	Architecture discovery and RCA

Python dependencies

Declared in pyproject.toml:

anthropic, boto3, httpx, jinja2, keyring, mcp, openai,
pydantic, python-dotenv, questionary, requests, rich, typer

Dev dependencies: pytest, pytest-asyncio, pytest-mock

Install dependencies

# End users (PyPI)
pipx install autoops-ai

# Contributors (from source)
pip install -e ".[dev]"

Example configuration

Copy the example env file — never commit real keys:

cp .env.example .env

See .env.example for all supported environment variables.

---

Install

From PyPI (recommended)

pipx install autoops-ai
# or
pip install autoops-ai

From source

git clone https://github.com/kenzzhood/AutoOps.git
cd AutoOps
pip install -e ".[dev]"

Install scripts

# macOS / Linux
curl -fsSL https://raw.githubusercontent.com/kenzzhood/AutoOps/main/scripts/install.sh | bash

# Windows PowerShell
pip install autoops-ai
# or run scripts/install.ps1 from the repo

---

Quick Start

# 1. Connect your AI provider (credentials saved to OS keychain)
autoops setup

# 2. Configure observability for any project
cd /path/to/your/app
autoops configure --repo .

# 3. Verify the stack
autoops doctor
autoops telemetry test

# 4. Investigate when an alert fires
autoops investigate --alert autoops_your-app_checkout_error_rate --window 30m

---

How It Works

When you run autoops configure --repo ., AutoOps executes six phases automatically:

Phase	What happens
Discovery	Scans repo files; LLM maps services, APIs, databases, critical paths → ~/.autoops/architecture.json
Instrumentation	Generates middleware and DB tracing in your codebase
Splunk Config	Creates dashboards, saved searches, alerts (autoops_{app}_*)
Evidence	Queries Splunk logs and metrics during investigation
RCA	LLM analyzes evidence and identifies root causes
Remediation	LLM recommends fixes; saves incident report to ~/.autoops/incidents/

Infrastructure bootstrapped in Docker:

Container	Ports	Role
autoops-splunk	8000 UI, 8088 HEC, 8089 REST	Log storage, dashboards, alerts
autoops-otel-collector	4317 gRPC, 4318 HTTP	Traces/metrics → Splunk HEC

Telemetry flows from your app to Splunk via HEC (JSON logs) and OpenTelemetry.

Full integration guide: docs/ARCHITECTURE.md

---

Command Reference

Command	Description
autoops setup	Interactive LLM provider setup
autoops configure --repo .	Full project observability setup
autoops doctor	Health check (Docker, Splunk, HEC, ports, LLM)
autoops provider list|set|test|show	Manage LLM profiles
autoops splunk start|status|stop|logs|open	Manage Splunk container
autoops telemetry start|status|test	Manage OTel Collector
autoops dashboards apply|list|open	Splunk dashboards
autoops alerts apply|list|test	Splunk alerts
autoops scan --repo .	Scan repo file tree
autoops instrument --repo .	Generate instrumentation only
autoops investigate --alert <name>	Run incident investigation
autoops watch --port 9000	Webhook listener for Splunk alerts
autoops demo start|bug-on|traffic	Local demo workflow

autoops init --repo <path> is an alias for configure.

---

Configuration

Environment variables

Copy .env.example to .env for optional env-var fallbacks when keyring is unavailable:

cp .env.example .env
# Edit .env with your provider keys (never commit .env)

Local state (~/.autoops/)

File	Contents
state.json	Splunk credentials, HEC token, dashboard/alert names
architecture.json	Discovered services, endpoints, critical paths
incidents/	Investigation reports (evidence + RCA + remediation)
config.json	LLM profile metadata (keys in keychain)

Splunk defaults created

Dashboards: Service Health Overview, Database Performance, Deployment Timeline, Incident Investigation, per-service health

Saved searches: error rate, latency, p95 latency, DB latency, deployments, no-data detection

Alerts: per-service error rate, checkout error rate, 5xx spike, ingestion stopped

Sample SPL:

index=main sourcetype=autoops | stats count by service, path, level
index=main sourcetype=autoops path="/checkout" | timechart avg(duration_ms) p95(duration_ms)

LLM providers

OpenAI · Claude (Anthropic) · Azure OpenAI · OpenRouter · Amazon Bedrock

Credentials are stored in your OS keychain, not in plain text.

---

Example Apps & Validation

Demo app (simple FastAPI)

Location: demo-app/

autoops demo start
cd demo-app && uvicorn main:app --port 8080
autoops demo bug-on
autoops demo traffic --requests 30
autoops investigate --alert checkout_error_rate

ShopVerse validation suite (microservices)

Location: validation/shopverse-platform/

7 microservices, PostgreSQL, Redis, injectable incident types, Docker Compose.

bash validation/reset_fresh.sh          # optional fresh start
pip install -e .
python3 scripts/seed_provider_from_env.py
cd validation/shopverse-platform && docker compose up -d --build
autoops configure --repo validation/shopverse-platform
python3 validation/run_validation.py

See validation/README.md and validation/validation_report.md.

Example architecture fixture

tests/fixtures/sample_architecture.json — sample discovered architecture for tests.

---

Development

git clone https://github.com/kenzzhood/AutoOps.git
cd AutoOps
pip install -e ".[dev]"
pytest

See CONTRIBUTING.md for the full contribution guide.

---

Built With

Python · Splunk Enterprise · OpenTelemetry · Docker · FastAPI · PostgreSQL · Redis · Typer · Pydantic · LLMs (OpenAI, Claude, Azure OpenAI, OpenRouter, Bedrock)

Full list: docs/BUILT_WITH.md

---

License

MIT — see LICENSE