Autonomous incident detection, AI-powered analysis, and self-healing code generation.
Zero human intervention from alert to pull request.
DevOps teams are drowning:
- 500+ alerts daily - Most are noise, but critical ones hide in the chaos
- Hours spent diagnosing - Manually correlating logs, metrics, and traces
- Repetitive fixes - The same issues require the same solutions
- Human bottleneck - Engineers are the single point of failure at 3 AM
What if infrastructure could heal itself?
An autonomous agent that:
- 🔍 Detects infrastructure incidents in real-time
- 🧠 Analyzes root causes using AI-powered log analysis
- 🔧 Generates code fixes automatically via agentic workflows
- 📝 Submits pull requests for review
- ✅ Validates changes with AI-powered code review
┌─────────────┐ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐
│ Incident │────▶│ Kestra │────▶│ Cline MCP │────▶│ GitHub PR │
│ Detection │ │ Orchestrator│ │ Auto-Fix │ │ + Review │
└─────────────┘ └──────┬──────┘ └─────────────┘ └─────────────┘
│
┌─────▼─────┐
│ AI │
│ Analysis │
└───────────┘
This project targets $15,000 across 5 sponsor tracks:
| Sponsor | Prize | Integration |
|---|---|---|
| Kestra | $4,000 | Workflow orchestration with AI Agent plugin |
| Cline | $5,000 | Custom MCP Server for autonomous code generation |
| Oumi | $3,000 | Fine-tuned SRE-LLM for log analysis |
| Vercel | $2,000 | Real-time mission control dashboard |
| CodeRabbit | $1,000 | AI-powered PR reviews of generated fixes |
┌──────────────────────────────────────────┐
│ SELF-HEALING PIPELINE │
└──────────────────────────────────────────┘
┌─────────────┐ ┌─────────────────────────────────────────────────────┐
│ │ │ │
│ Log Source │────────▶│ ┌─────────┐ ┌─────────┐ ┌──────────────┐ │
│ (Webhook) │ │ │ Kestra │───▶│ AI │───▶│ Severity │ │
│ │ │ │ Trigger │ │ Analysis│ │ Router │ │
└─────────────┘ │ └─────────┘ └─────────┘ └──────┬───────┘ │
│ │ │
│ ┌──────────────────────────┴───┐ │
│ ▼ ▼ │
│ ┌─────────────────┐ ┌────────────────┐ │
│ │ HIGH/CRITICAL │ │ LOW/MEDIUM │ │
│ │ ┌───────────┐ │ │ │ │
│ │ │ MCP Server│ │ │ Log & Alert │ │
│ │ │ (Cline) │ │ │ │ │
│ │ └─────┬─────┘ │ └────────────────┘ │
│ │ │ │ │
│ │ ┌─────▼─────┐ │ │
│ │ │ Generate │ │ │
│ │ │ Fix │ │ │
│ │ └─────┬─────┘ │ │
│ │ │ │ │
│ │ ┌─────▼─────┐ │ │
│ │ │ Create PR │ │ │
│ │ └─────┬─────┘ │ │
│ └────────┼────────┘ │
│ │ │
└───────────┼───────────────────────────────────────┘
│
▼
┌───────────────────────┐
│ CodeRabbit │
│ AI Code Review │
└───────────────────────┘
│
▼
┌───────────────────────┐
│ ✅ Merge Ready │
└───────────────────────┘
┌─────────────────────────────────────────────────────────────────────────┐
│ REAL-TIME DASHBOARD │
│ (Vercel - Next.js) │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ ┌─────────────┐ │
│ │System Health │ │ Incidents │ │ Activity Feed│ │ Metrics │ │
│ └──────────────┘ └──────────────┘ └──────────────┘ └─────────────┘ │
└─────────────────────────────────────────────────────────────────────────┘
- Webhook-based ingestion from any monitoring tool
- Real-time log analysis and classification
- Severity-based routing (Critical → Auto-fix, Low → Log only)
- Custom SRE-LLM trained on infrastructure patterns
- Root cause identification
- Suggested remediation with confidence scores
- Cline MCP Server integration for autonomous coding
- Context-aware fixes based on repository structure
- Automatic PR creation with detailed descriptions
- Live incident tracking
- Activity stream with SSE updates
- System health monitoring
- Demo controls for testing
- CodeRabbit integration for automated PR reviews
- AI reviewing AI-generated code
- Quality gates before merge
- Node.js 18+
- Docker & Docker Compose
- Git
- GitHub Account + Personal Access Token
- OpenAI API Key
# Clone the repository
git clone https://github.com/me1abu/devOpsOrchestrate.git
cd devOpsOrchestrate
# Copy environment variables
cp .env.example .env
# Edit .env with your API keys
# Start the infrastructure
docker-compose up -d
# Start the dashboard (development)
cd dashboard
npm install
npm run dev| Service | URL | Credentials |
|---|---|---|
| Dashboard | http://localhost:3000 | - |
| Kestra UI | http://localhost:8080 | admin@kestra.io / Kestra123 |
| MCP Server | http://localhost:3001 | - |
devOpsOrchestrate/
├── 📂 dashboard/ # Next.js real-time dashboard
│ ├── app/ # App router pages
│ ├── components/ # React components
│ └── Dockerfile
│
├── 📂 kestra/flows/ # Kestra workflow definitions
│ ├── main-orchestrator.yml # Main incident processing flow
│ └── auto-fix-workflow.yml # Autonomous remediation flow
│
├── 📂 mcp-server/ # Cline MCP Server
│ ├── src/
│ │ ├── index.ts # Express server + SSE
│ │ └── tools.ts # MCP tool definitions
│ └── Dockerfile
│
├── 📂 oumi/ # Oumi model training
│ ├── data/ # Training data (JSONL)
│ ├── train.py # Training script
│ └── config.yaml # Model configuration
│
├── 📂 monitoring/ # Demo monitoring setup
├── 📂 scripts/ # Utility scripts
├── 🐳 docker-compose.yml # Full stack deployment
└── 📄 README.md
The MCP Server exposes these tools for Cline integration:
| Tool | Description |
|---|---|
get_pending_incidents() |
Fetch unresolved incidents |
get_incident_details(id) |
Get full incident context |
get_repository_context() |
Understand codebase structure |
apply_fix(incident_id, fix) |
Apply generated fix |
create_pull_request(...) |
Create GitHub PR |
report_fix_status(...) |
Update incident status |
GET / # API documentation
GET /health # Health check
GET /events # SSE stream for real-time updates
GET /incidents # List all incidents
POST /incidents # Create new incident
GET /incidents/:id # Get incident details
PATCH /incidents/:id # Update incident
GET /stats # Get statistics
- Dashboard: autosre.vercel.app
- MCP Server: mcp-server-deploy.up.railway.app
📺 Watch the 3-minute demo (Coming soon)
curl -X POST https://mcp-server-deploy.up.railway.app/incidents \
-H "Content-Type: application/json" \
-d '{
"severity": "critical",
"category": "database",
"summary": "Connection pool exhausted",
"description": "FATAL: max_connections=100 exceeded",
"source": "postgresql"
}'# Example: Main orchestrator flow
id: self-healing-orchestrator
namespace: devops.healing
tasks:
- id: analyze-incident
type: io.kestra.plugin.scripts.python.Script
script: |
# AI-powered log analysis
# Severity classification
# Root cause identification
- id: trigger-autofix
type: io.kestra.plugin.core.flow.If
condition: "{{ severity == 'critical' }}"
then:
- id: call-mcp-server
type: io.kestra.plugin.core.http.Request
uri: "{{ mcp_server_url }}/fix"// MCP tools for autonomous code generation
const tools = [
{
name: "get_pending_incidents",
description: "Fetch incidents awaiting fixes",
handler: async () => await db.getIncidents({ status: "pending" })
},
{
name: "create_pull_request",
description: "Create a GitHub PR with the fix",
handler: async ({ title, body, branch }) => {
return await github.createPR({ title, body, branch });
}
}
];# Training data format
{
"input": "FATAL: Connection pool exhausted - max_connections=100 exceeded",
"output": {
"severity": "critical",
"category": "database",
"root_cause": "Connection pool limit reached",
"suggested_fix": "Increase max_connections or implement connection pooling"
}
}- Next.js 14 with App Router
- Real-time updates via Server-Sent Events
- Responsive design with Tailwind CSS
# .coderabbit.yaml
reviews:
auto_review:
enabled: true
path_filters:
- "!**/*.md"
tools:
github-checks:
enabled: true- Multi-cloud support - AWS, GCP, Azure integrations
- Slack/PagerDuty integration - Alert routing
- Learning from feedback - Improve fixes based on PR reviews
- Rollback automation - Auto-revert failed deployments
- Cost optimization - Infrastructure right-sizing recommendations
Contributions are welcome! Please read our Contributing Guidelines first.
# Fork the repo
# Create your feature branch
git checkout -b feature/amazing-feature
# Commit your changes
git commit -m 'Add amazing feature'
# Push to the branch
git push origin feature/amazing-feature
# Open a Pull RequestMIT License - see LICENSE for details.
Abu
- GitHub: @me1abu