Add resource limits, abuse protection and overload handling

Running games always continue; new work is refused when there is no
capacity; abuse is rejected immediately.

- Docker: CPU/memory/pids caps per container (overridable in .env);
  uvicorn message size, queue and concurrency limits
- Server abuse limits: connections total/per IP (close 1013), message
  rate per connection (token bucket, close 1008 on floods), lobby
  creation rate per IP
- Load monitor: event loop lag, CPU of the own cgroup quota and
  throttling, with hysteresis (normal/busy/overloaded, severe lag
  overloads immediately)
- Admission control: create_lobby and start_game refused while
  overloaded or MAX_ACTIVE_GAMES are running; server_status message
  broadcast on changes; /health reports load but stays ok
- AI bridge: timeout sent with requests, in-flight cap, circuit breaker
- AI service: bounded workers and queue ("overloaded"), drops expired
  requests, degrades to the heuristic agent when busy or too slow
- Website: busy notices, Create/Start disabled when not accepting
  games, slow retry when the server is full
- Tests for limiters, load levels, admission control and AI service
  limits; docs updated

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

compose.yaml

@@ -4,6 +4,12 @@ services:
   wizard_website:
     build: ./wizard_website
     restart: unless-stopped
+    deploy:
+      resources:
+        limits:
+          cpus: "${WEBSITE_CPUS:-0.25}"
+          memory: ${WEBSITE_MEMORY:-64M}
+    pids_limit: 64
     networks:
       - wizard_website
 
@@ -14,6 +20,16 @@ services:
       AI_SHARED_SECRET: ${AI_SHARED_SECRET:?set AI_SHARED_SECRET in .env}
       ALLOWED_ORIGINS: ${ALLOWED_ORIGINS:-}
       LOG_LEVEL: ${LOG_LEVEL:-INFO}
+      MAX_ACTIVE_GAMES: ${MAX_ACTIVE_GAMES:-50}
+      MAX_CONNECTIONS: ${MAX_CONNECTIONS:-500}
+      MAX_CONNECTIONS_PER_IP: ${MAX_CONNECTIONS_PER_IP:-10}
+    # Hard caps so Wizard can never starve other services on the host.
+    deploy:
+      resources:
+        limits:
+          cpus: "${SERVER_CPUS:-1.0}"
+          memory: ${SERVER_MEMORY:-512M}
+    pids_limit: 256
     networks:
       - wizard_api
       - wizard_internal
@@ -31,6 +47,18 @@ services:
       AGENT: ${AI_AGENT:-model}
       MODEL_PATH: /srv/models/latest
       LOG_LEVEL: ${LOG_LEVEL:-INFO}
+      AI_WORKERS: ${AI_WORKERS:-1}
+      AI_MAX_QUEUE: ${AI_MAX_QUEUE:-64}
+      AI_DEGRADE_QUEUE: ${AI_DEGRADE_QUEUE:-8}
+      # Numeric libraries (numpy/torch) must not start one thread per host core.
+      OMP_NUM_THREADS: ${AI_WORKERS:-1}
+      MKL_NUM_THREADS: ${AI_WORKERS:-1}
+    deploy:
+      resources:
+        limits:
+          cpus: "${AI_CPUS:-1.0}"
+          memory: ${AI_MEMORY:-1G}
+    pids_limit: 256
     volumes:
       # Trained checkpoints; replace without rebuilding the image.
       - ./wizard_ai/models:/srv/models:ro