Fix compose: move pids limits into deploy.resources.limits

Newer docker compose treats pids_limit and deploy.resources.limits.pids
as the same setting and rejects the project when both are present
("can't set distinct values on 'pids_limit' and
'deploy.resources.limits.pids'"). The limits are now only set as
deploy.resources.limits.pids (same values).

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

compose.yaml

@@ -9,7 +9,7 @@ services:
         limits:
           cpus: "${WEBSITE_CPUS:-0.25}"
           memory: ${WEBSITE_MEMORY:-64M}
-    pids_limit: 64
+          pids: 64
     networks:
       - wizard_website
 
@@ -29,7 +29,7 @@ services:
         limits:
           cpus: "${SERVER_CPUS:-1.0}"
           memory: ${SERVER_MEMORY:-512M}
-    pids_limit: 256
+          pids: 256
     networks:
       - wizard_api
       - wizard_internal
@@ -58,7 +58,7 @@ services:
         limits:
           cpus: "${AI_CPUS:-1.0}"
           memory: ${AI_MEMORY:-1G}
-    pids_limit: 256
+          pids: 256
     volumes:
       # Trained checkpoints; replace without rebuilding the image.
       - ./wizard_ai/models:/srv/models:ro

docs/09-limits-and-performance.md

@@ -60,7 +60,7 @@ Set in `compose.yaml` from `.env` (defaults shown):
   would end all games (state is in memory), so the limit has lots of headroom:
   the server uses ~80 MB idle and ~100 MB with 100 games (see
   [measurements](#measurements)).
-- **Processes** (`pids_limit`) stop fork bombs / thread explosions.
+- **Processes** (`deploy.resources.limits.pids`) stop fork bombs / thread explosions.
 - The AI container sets `OMP_NUM_THREADS` / `MKL_NUM_THREADS` to `AI_WORKERS`,
   so numeric libraries (numpy, torch) don't start one thread per *host* core
   inside a 1-CPU container.