Configuration¶

Orion uses environment variables for all configuration across every layer of the stack — the Go gateway, Python services, dashboard, and infrastructure. Copy .env.example to .env and adjust values as needed. In Docker Compose, variables are loaded automatically from the .env file at the project root.

# Create your local configuration
cp .env.example .env

# Edit with your preferred editor
$EDITOR .env

Gateway Configuration¶

The Go gateway reads configuration from environment variables via pkg/config/config.go:

openssl rand -base64 64

Service URLs¶

The gateway proxies requests to Python services using these URLs. Each variable maps to one of the six downstream microservices:

In Docker Compose, these automatically resolve to container hostnames:

# Docker Compose (default in deploy/docker-compose.yml)
SCOUT_URL=http://scout:8001
DIRECTOR_URL=http://director:8002
MEDIA_URL=http://media:8003
EDITOR_URL=http://editor:8004
PULSE_URL=http://pulse:8005
PUBLISHER_URL=http://publisher:8006
IDENTITY_URL=http://identity:8007

When running services locally (without Docker), use localhost with the appropriate port.

OAuth & Authentication¶

The gateway supports OAuth 2.0 login with GitHub and Google. Configure these to enable social login:

Email (SMTP)¶

The Identity service sends transactional emails for verification, password reset, and user invitations. In development, emails are logged to the console.

Data Layer¶

Shared across all Python services via orion_common.config.CommonSettings:

Verify data layer connectivity¶

# PostgreSQL
psql -h localhost -U orion -d orion -c "SELECT 1;"

# Redis
redis-cli -u redis://localhost:6379 ping

# Milvus
curl http://localhost:9091/healthz

AI Infrastructure¶

Verify AI infrastructure¶

# Check Ollama is running and list available models
curl http://localhost:11434/api/tags

# Check ComfyUI is accessible
curl http://localhost:8188/system_stats

Switch between LOCAL and CLOUD providers at runtime¶

# Switch Media to Fal.ai cloud
orion provider switch media --mode CLOUD --provider fal_ai

# Switch back to local ComfyUI
orion provider switch media --mode LOCAL --provider comfyui

# View all active providers and their status
orion provider status

Dashboard¶

The dashboard reads this at build time (Next.js public env var). For development, the default works out of the box.

Python Service Settings¶

Each Python service inherits from CommonSettings in libs/orion-common/. Settings are loaded via Pydantic's BaseSettings with automatic environment variable binding:

from orion_common.config import get_settings

settings = get_settings()  # Cached singleton
print(settings.database_url)  # postgresql+asyncpg://orion:orion_dev@localhost:5432/orion

The database_url property auto-constructs the async connection string from individual POSTGRES_* variables. A database_url_sync variant exists for Alembic migrations.

Override settings per service¶

Individual services can define additional settings by extending CommonSettings:

from orion_common.config import CommonSettings

class ScoutSettings(CommonSettings):
    scan_interval_minutes: int = 30
    min_trend_score: float = 0.5
    max_trends_per_scan: int = 50

:material-rate-limit: Rate Limiting¶

Per-service rate limits are configured in the gateway router. Limits use Redis-backed sliding windows to ensure consistency across gateway instances.

Rate limit response headers¶

Every response from the gateway includes rate limit metadata:

Example rate-limited response¶

curl -i http://localhost:8000/api/v1/scout/api/v1/trends
# HTTP/1.1 429 Too Many Requests
# X-RateLimit-Limit: 10
# X-RateLimit-Remaining: 0
# X-RateLimit-Reset: 1710345600
# Retry-After: 45