Files
psyc/docker-compose.yml
m17hr1l b51a88d502 stage-26b: Docker topology in /admin — read-only socket-proxy + graph
New tecnativa/docker-socket-proxy sidecar exposes only GET on
containers/networks/info/ping; POST and DELETE are blocked. The cockpit
queries it over the backend network — /var/run/docker.sock is never
mounted into a web-facing container.

cockpit/docker_view.py normalizes the daemon view: containers carry
per-network IP/MAC + published_ports; networks carry subnet/gateway from
IPAM; host_info pulls /info (degrades gracefully). topology() returns
the combined snapshot.

/admin/docker (admin-gated): a force-directed graph (pure SVG +
vanilla JS, ~280 lines) renders the complete setup — a host node,
switch nodes with subnet labels colored by driver, container nodes
colored by state, member wires labeled with the container's IP on that
network, uplinks from non-internal switches to the host labeled with
the gateway, and dashed publish-edges from containers to the host for
their published ports. Drag to rearrange, scroll to zoom, re-settle
kicks the physics. Below the graph: containers table + grouped network
cards as a textual mirror. 12 docker_view tests; verified live (32
containers, 11 switches, real subnets + gateways).

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-23 03:08:39 +02:00

99 lines
3.3 KiB
YAML

# psyc — neuronetz.ai deployment stack.
#
# docker compose up -d --build # cockpit + mock-cert (no GPU)
# docker compose --profile gpu up -d --build # + the live model (needs an NVIDIA GPU)
#
# The cockpit is fronted by the external `backend` network's nginx-proxy as
# psyc.neuronetz.ai (point DNS for that name at the proxy host). mock-cert and
# the inference server stay internal — no VIRTUAL_HOST, reachable only inside
# `backend` by service name.
#
# WARNING: psyc has no built-in authentication. The reverse proxy / network
# perimeter is the security boundary. See docs/deploy.md.
services:
cockpit:
build: .
image: psyc:latest
command: ["psyc", "serve", "--host", "0.0.0.0", "--port", "8767"]
env_file: .env # per-dev API keys (gitignored). cp .env.example .env first.
environment:
VIRTUAL_HOST: psyc.neuronetz.ai
VIRTUAL_PORT: "8767"
PSYC_MOCK_CERT_URL: http://mock-cert:8770
PSYC_SOAR_URL: http://mock-cert:8770
PSYC_INFERENCE_URL: http://inference:8771
PSYC_DOCKER_PROXY: http://docker-socket-proxy:2375
ports:
- "8767:8767" # direct/debug access; the proxy serves psyc.neuronetz.ai on :80
volumes:
- ./data:/data
networks: [backend]
restart: unless-stopped
healthcheck:
test: ["CMD", "python", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:8767/healthz')"]
interval: 30s
timeout: 5s
retries: 3
mock-cert:
image: psyc:latest
command: ["psyc", "mock-cert", "--host", "0.0.0.0", "--port", "8770"]
volumes:
- ./data:/data
networks: [backend]
restart: unless-stopped
healthcheck:
test: ["CMD", "python", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:8770/healthz')"]
interval: 30s
timeout: 5s
retries: 3
# Read-only Docker daemon proxy. The cockpit's /admin/docker view queries this
# over the backend network instead of touching /var/run/docker.sock directly,
# so a compromise of the web app can't drive the daemon. Only GET on
# containers/networks/ping is enabled — POST/DELETE/EXEC stay blocked.
docker-socket-proxy:
image: tecnativa/docker-socket-proxy:0.3
environment:
CONTAINERS: "1"
NETWORKS: "1"
PING: "1"
INFO: "1"
POST: "0"
DELETE: "0"
volumes:
- /var/run/docker.sock:/var/run/docker.sock:ro
networks: [backend]
restart: unless-stopped
# The live fine-tuned model behind the Classifier bot. GPU-only — opt in with
# `--profile gpu`. Uses the psyc-trainer image (built from Dockerfile.train).
inference:
image: psyc-trainer
command: ["/scripts/serve_model.py", "--adapter", "/data/adapters/psyc-v5/final", "--host", "0.0.0.0", "--port", "8771"]
volumes:
- ./data:/data
- ./scripts:/scripts
networks: [backend]
restart: unless-stopped
profiles: ["gpu"]
healthcheck:
test: ["CMD", "python", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:8771/healthz')"]
interval: 30s
timeout: 5s
retries: 3
start_period: 90s
deploy:
resources:
reservations:
devices:
- driver: nvidia
count: all
capabilities: [gpu]
networks:
backend:
name: backend
external: true