feat(infra): add nginx WebSocket proxy and SSH relay config to compose files

- Add WebSocket upgrade map to nginx and proxy /ws/ssh to poller:8080
- Update CSP connect-src to allow ws: and wss: for terminal connections
- Add tunnel port range 49000-49100, SSH relay env vars, ulimits, and healthcheck to poller in both override and prod compose files
- Increase poller memory limit to 512M in prod for tunnel/SSH overhead

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

docker-compose.override.yml

@@ -81,6 +81,20 @@ services:
       OPENBAO_TOKEN: dev-openbao-token
       POLL_INTERVAL_SECONDS: 60
       WIREGUARD_GATEWAY: wireguard
+      TUNNEL_PORT_MIN: 49000
+      TUNNEL_PORT_MAX: 49100
+      TUNNEL_IDLE_TIMEOUT: 300
+      SSH_RELAY_PORT: 8080
+      SSH_IDLE_TIMEOUT: 900
+      SSH_MAX_SESSIONS: 200
+      SSH_MAX_PER_USER: 10
+      SSH_MAX_PER_DEVICE: 20
+    ports:
+      - "127.0.0.1:49000-49100:49000-49100"
+    ulimits:
+      nofile:
+        soft: 8192
+        hard: 8192
     cap_add:
       - NET_ADMIN
     depends_on:
@@ -90,6 +104,11 @@ services:
         condition: service_healthy
       nats:
         condition: service_healthy
+    healthcheck:
+      test: ["CMD-SHELL", "wget --spider -q http://localhost:8080/healthz || exit 1"]
+      interval: 30s
+      timeout: 3s
+      retries: 3
     deploy:
       resources:
         limits:

docker-compose.prod.yml

@@ -44,6 +44,20 @@ services:
     environment:
       ENVIRONMENT: production
       LOG_LEVEL: info
+      TUNNEL_PORT_MIN: 49000
+      TUNNEL_PORT_MAX: 49100
+      TUNNEL_IDLE_TIMEOUT: 300
+      SSH_RELAY_PORT: 8080
+      SSH_IDLE_TIMEOUT: 900
+      SSH_MAX_SESSIONS: 200
+      SSH_MAX_PER_USER: 10
+      SSH_MAX_PER_DEVICE: 20
+    ports:
+      - "127.0.0.1:49000-49100:49000-49100"
+    ulimits:
+      nofile:
+        soft: 8192
+        hard: 8192
     depends_on:
       postgres:
         condition: service_healthy
@@ -51,10 +65,15 @@ services:
         condition: service_healthy
       nats:
         condition: service_healthy
+    healthcheck:
+      test: ["CMD-SHELL", "wget --spider -q http://localhost:8080/healthz || exit 1"]
+      interval: 30s
+      timeout: 3s
+      retries: 3
     deploy:
       resources:
         limits:
-          memory: 256M
+          memory: 512M  # increased from 256M for tunnel/SSH overhead
     restart: unless-stopped
     logging:
       driver: json-file

infrastructure/docker/nginx-spa.conf

@@ -1,3 +1,8 @@
+map $http_upgrade $connection_upgrade {
+    default upgrade;
+    '' close;
+}
+
 server {
     listen 80;
     server_name _;
@@ -18,7 +23,7 @@ server {
 
     # CSP for React SPA with Tailwind CSS and Leaflet maps
     # worker-src required for SRP key derivation Web Worker (Safari won't fall back to script-src)
-    add_header Content-Security-Policy "default-src 'self'; script-src 'self'; style-src 'self' 'unsafe-inline'; img-src 'self' data: https://*.tile.openstreetmap.org; font-src 'self'; connect-src 'self'; worker-src 'self'; frame-ancestors 'none'; base-uri 'self'; form-action 'self';" always;
+    add_header Content-Security-Policy "default-src 'self'; script-src 'self'; style-src 'self' 'unsafe-inline'; img-src 'self' data: https://*.tile.openstreetmap.org; font-src 'self'; connect-src 'self' ws: wss:; worker-src 'self'; frame-ancestors 'none'; base-uri 'self'; form-action 'self';" always;
 
     # Proxy API requests to the backend service
     # The api container is reachable via Docker internal DNS as "api" on port 8000
@@ -40,6 +45,29 @@ server {
         proxy_hide_header Content-Security-Policy;
     }
 
+    # WebSocket proxy for SSH terminal
+    location /ws/ssh {
+        resolver 127.0.0.11 valid=10s ipv6=off;
+        set $poller_upstream http://poller:8080;
+
+        proxy_pass $poller_upstream;
+        proxy_http_version 1.1;
+
+        proxy_set_header Upgrade $http_upgrade;
+        proxy_set_header Connection $connection_upgrade;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header Host $host;
+
+        proxy_read_timeout 1800s;
+        proxy_send_timeout 1800s;
+
+        proxy_buffering off;
+        proxy_request_buffering off;
+        proxy_busy_buffers_size 512k;
+        proxy_buffers 8 512k;
+    }
+
     # Serve static assets with long cache headers
     # Note: add_header in a location block clears parent-block headers,
     # so we re-add the essential security header for static assets.