DevOps Sentinel Next

Agent-accessible SRE operations platform for health checks, incident operations, evidence-backed response plans, and postmortems.

This is next-generation distribution of DevOps Sentinel. It keeps Python as canonical runtime while adding a secure MCP interface and a typed npm client.

Why use it?

• Fast signal: check one or many endpoints with latency, status, SSL, and suggestions.
• Incident memory: preserve incidents, event timelines, response plans, and postmortems.
• Agent-ready: expose operational context to Claude, Cursor, and other MCP hosts.
• API-first: CLI, web console, MCP, and npm client use shared contracts.
• Safe by default: remediation is not exposed as an unapproved shell; destructive runbook steps require approval.
• Self-hostable: Supabase persistence, optional AI, no telemetry, MIT licensed.

Architecture

CLI / FastAPI / MCP / npm client
            ↓
     Sentinel application
            ↓
   Supabase + optional AI

Install

pip install devops-sentinel-next
# Optional MCP server
pip install "devops-sentinel-next[mcp]"

The CLI command remains sentinel:

sentinel doctor
sentinel health https://api.example.com/health
sentinel services list

MCP

Run local stdio server:

devops-sentinel-mcp

Read-only tools include health_check, health_check_batch, doctor, list_incidents, get_incident, get_incident_events, analyze_anomaly, and generate_postmortem. Account-scoped tools require CLI authentication and Supabase configuration. Remote MCP deployment needs an authentication and audit layer; do not expose it directly to the public internet.

npm client

A separate package lives in packages/client:

cd packages/client
npm install
npm run build

Publish target: @devops-sentinel/client. It calls the HTTP API; it does not duplicate Python monitoring logic.

How it works

1. Register service endpoint.
2. Run health checks on schedule or on demand.
3. Store health-check evidence and latest service state.
4. Open incidents after failure policy evaluation.
5. Record detection, investigation, resolution, and postmortem events.
6. Generate fallback or optional AI-assisted postmortems.
7. Let humans approve any destructive remediation.

Configuration

SUPABASE_URL=https://your-project.supabase.co
SUPABASE_ANON_KEY=your_key
SENTINEL_WEB_URL=https://your-console.example.com
LLM_PROVIDER=openrouter
OPENROUTER_API_KEY=your_key

Run sentinel doctor before operating.

Security and privacy

• No telemetry or external data collection.
• Data stays in your Supabase project.
• Credentials are stored under ~/.sentinel.
• Destructive runbook actions require approval.
• Arbitrary commands and script checks are disabled unless explicitly enabled.
• Review SSRF, authentication, and authorization controls before public deployment.

New PyPI package release

This repository now builds distribution devops-sentinel-next; import path and CLI remain compatible. Create or use new PyPI account, verify name ownership, then:

python -m pip install --upgrade build twine
python -m build
python -m twine check dist/*
python -m twine upload dist/*

Use a PyPI API token, not account password. Configure trusted publishing from GitHub Actions before automation.

Development

python -m pip install -e ".[dev,mcp]"
pytest -q -o addopts=""
python -m ruff check sentinel tests

MIT License.