70 lines
1.9 KiB
YAML
70 lines
1.9 KiB
YAML
services:
|
|
ai-orchestrator:
|
|
build:
|
|
context: .
|
|
dockerfile: Dockerfile
|
|
image: ai-orchestrator:latest
|
|
container_name: ai-orchestrator
|
|
restart: unless-stopped
|
|
environment:
|
|
- RUNPOD_API_KEY=${RUNPOD_API_KEY}
|
|
- OLLAMA_HOST=http://ollama:11434
|
|
depends_on:
|
|
ollama:
|
|
condition: service_started
|
|
labels:
|
|
# Traefik auto-discovery
|
|
- "traefik.enable=true"
|
|
- "traefik.http.routers.ai-orchestrator.rule=Host(`ai.jeffemmett.com`)"
|
|
- "traefik.http.routers.ai-orchestrator.entrypoints=websecure"
|
|
- "traefik.http.routers.ai-orchestrator.tls=true"
|
|
- "traefik.http.services.ai-orchestrator.loadbalancer.server.port=8080"
|
|
# Health check for Traefik
|
|
- "traefik.http.services.ai-orchestrator.loadbalancer.healthcheck.path=/api/health"
|
|
- "traefik.http.services.ai-orchestrator.loadbalancer.healthcheck.interval=30s"
|
|
networks:
|
|
- traefik-public
|
|
- ai-internal
|
|
healthcheck:
|
|
test: ["CMD", "python", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:8080/api/health')"]
|
|
interval: 30s
|
|
timeout: 10s
|
|
retries: 3
|
|
start_period: 10s
|
|
|
|
ollama:
|
|
image: ollama/ollama:latest
|
|
container_name: ollama
|
|
restart: unless-stopped
|
|
volumes:
|
|
- ollama-data:/root/.ollama
|
|
networks:
|
|
- ai-internal
|
|
# Expose internally only (orchestrator routes to it)
|
|
expose:
|
|
- "11434"
|
|
# Note: Ollama doesn't have curl/wget, so we use a simple TCP check
|
|
healthcheck:
|
|
test: ["CMD-SHELL", "ollama list || exit 0"]
|
|
interval: 60s
|
|
timeout: 30s
|
|
retries: 3
|
|
start_period: 60s
|
|
# CPU-only mode (no GPU passthrough needed for RS 8000)
|
|
deploy:
|
|
resources:
|
|
limits:
|
|
memory: 16G
|
|
reservations:
|
|
memory: 4G
|
|
|
|
volumes:
|
|
ollama-data:
|
|
driver: local
|
|
|
|
networks:
|
|
traefik-public:
|
|
external: true
|
|
ai-internal:
|
|
driver: bridge
|