Configuration Reference

AIA uses aia.config.yaml as the main configuration file, with support for local overrides via aia.config.local.yaml.

#Configuration Files

Main Config: `aia.config.yaml`

The primary configuration file (committed to git):

aia.config.yaml

project_name: "aia-demo"
cluster_env: "prod"

services:
  dashboard: { port: 3000, base_url: "http://localhost:3000" }
  router: { port: 3001, base_url: "http://localhost:3001" }
  autopsy: { port: 3002, base_url: "http://localhost:3002" }
  state: { port: 3003, base_url: "http://localhost:3003" }
  git: { port: 3004, base_url: "http://localhost:3004" }
  repro: { port: 3005, base_url: "http://localhost:3005" }
  web: { port: 3006, base_url: "http://localhost:3006" }
  docs: { port: 3007, base_url: "http://localhost:3007" }
  sample_app: { port: 3008, base_url: "http://localhost:3008" }
  agent: { port: 4318, base_url: "http://localhost:4318" }

github:
  provider: "github"
  token: "PLACEHOLDER"  # Read from .env
  owner: "your-org"  # CHANGE THIS
  repo: "your-repo"  # CHANGE THIS
  base_branch: "main"

ai:
  provider: "you.com"
  model: "express"
  api_url: "https://api.you.com/v1/chat/completions"

Local Override: `aia.config.local.yaml`

For local testing (gitignored):

aia.config.local.yaml

github:
  owner: "your-github-username"
  repo: "your-test-repo"

Priority: aia.config.local.yaml > aia.config.yaml

#Environment Variables

Create a .env file in the project root:

.env

# Required
YOU_API_KEY=ydc_your_actual_key_here
GITHUB_TOKEN=ghp_your_actual_token_here
DATABASE_URL=postgresql://user:pass@host/db

# Optional: R2 Storage
R2_ACCESS_KEY_ID=your_r2_access_key
R2_SECRET_ACCESS_KEY=your_r2_secret_key
R2_BUCKET_NAME=aia
R2_ACCOUNT_ID=your_account_id

#Configuration Schema

Top-Level Fields

Services

Each service has:

port: Port number
base_url: Full URL with protocol

Terminal

services:
  # OTEL receiver
  agent: { port: 4318, base_url: "http://localhost:4318" }
  
  # Dashboard UI
  dashboard: { port: 3000, base_url: "http://localhost:3000" }
  
  # Incident orchestrator
  router: { port: 3001, base_url: "http://localhost:3001" }
  
  # AI analysis
  autopsy: { port: 3002, base_url: "http://localhost:3002" }
  
  # Database service
  state: { port: 3003, base_url: "http://localhost:3003" }
  
  # GitHub integration
  git: { port: 3004, base_url: "http://localhost:3004" }
  
  # Repro service (optional)
  repro: { port: 3005, base_url: "http://localhost:3005" }
  
  # Marketing site
  web: { port: 3006, base_url: "http://localhost:3006" }
  
  # Documentation
  docs: { port: 3007, base_url: "http://localhost:3007" }
  
  # Sample app for testing
  sample_app: { port: 3008, base_url: "http://localhost:3008" }

GitHub Configuration

Terminal

github:
  provider: "github"
  token: "PLACEHOLDER"  # Read from GITHUB_TOKEN env var
  owner: "your-github-username"  # REQUIRED
  repo: "your-repo-name"  # REQUIRED
  base_branch: "main"
  username: "aia-bot"  # Optional: bot username
  email: "bot@aia.local"  # Optional: bot email

AI Configuration

Terminal

ai:
  provider: "you.com"
  model: "express"  # or "research-pro"
  api_url: "https://api.you.com/v1/chat/completions"

Models:

express: Fast, good quality
research-pro: Slower, better quality

Paths Configuration

Terminal

paths:
  repo_root: "/path/to/project"
  logs: "logs/app.log"
  events: "events"
  storage: "storage"
  autopsy_output: "autopsy_output"
  patches: "patches"
  pr_description: "pr_descriptions"
  repro_logs: "repro_logs"

Storage Configuration (Optional)

R2 is optional for autopsy result backups:

Terminal

storage:
  provider: "r2"
  bucket: "aia"
  access_key: "${R2_ACCESS_KEY_ID}"
  secret_key: "${R2_SECRET_ACCESS_KEY}"
  account_id: "${R2_ACCOUNT_ID}"
  region: "auto"

#Environment Variable Reference

Required

Optional

#Config Loading

The config loader searches for files in this order:

$AIA_CONFIG_PATH (if set)
./aia.config.local.yaml
../aia.config.local.yaml
../../aia.config.local.yaml
../../../aia.config.local.yaml
./config/aia.config.local.yaml
./config/aia.config.yaml
./aia.config.yaml
../aia.config.yaml
../../aia.config.yaml
../../../aia.config.yaml

Why multiple paths? Services run from different directories (apps/git/, apps/agent/, etc.), so the loader searches upward to find the config in the project root.

#Database Configuration

PostgreSQL (Recommended)

Set in .env:

Terminal

DATABASE_URL=postgresql://user:password@host:5432/database

Providers:

Neon - Free tier available
Supabase - Free tier available
Local PostgreSQL

Schema

Tables are auto-created:

incidents: Incident metadata
autopsy_results: AI analysis results

#Production Configuration

Security Best Practices

Never commit secrets
- Use .env for all credentials
- Add .env to .gitignore
- Use aia.config.local.yaml for local testing
Use environment variables

Terminal
github: token: "${GITHUB_TOKEN}" # Read from env
Rotate tokens regularly
- GitHub PAT: Every 90 days
- You.com API key: As needed

Production Checklist

[ ] Set cluster_env: "prod"
[ ] Configure production database
[ ] Set up R2 storage (optional)
[ ] Use production GitHub repo
[ ] Set proper service URLs (not localhost)
[ ] Enable HTTPS for external services
[ ] Set up monitoring and logging

#Example Configurations

Development

aia.config.yaml (dev)

project_name: "aia-dev"
cluster_env: "dev"

github:
  owner: "your-username"
  repo: "test-repo"

ai:
  model: "express"  # Faster for dev

Production

aia.config.yaml (prod)

project_name: "aia-prod"
cluster_env: "prod"

services:
  dashboard: { port: 3000, base_url: "https://dashboard.example.com" }
  agent: { port: 4318, base_url: "https://agent.example.com" }
  # ... other services

github:
  owner: "your-org"
  repo: "production-app"

ai:
  model: "research-pro"  # Better quality for prod

#Troubleshooting

"Config file not found"

Check file exists: ls aia.config.yaml
Verify file name (not .yml)
Check search paths in logs

"Invalid config"

Validate YAML syntax
Check required fields are present
Verify service ports don't conflict

"GitHub token invalid"

Check token in .env
Verify token has repo scope
Test token: curl -H "Authorization: token $GITHUB_TOKEN" https://api.github.com/user

"Database connection failed"

Verify DATABASE_URL format
Check database is running
Test connection: psql $DATABASE_URL

#Next Steps

Installation - Setup guide
GitHub Integration - GitHub setup
Architecture - System overview

Installation Running Agent