Harden the container runtime and shut down cleanly

Running the stack on built output stopped the idle CPU load, but the
remaining findings from the deployment report were still open.

The Dockerfile is now multi-stage. The prod target runs on node:24-slim as
the "node" user instead of root and carries only its runtime dependencies;
"npm prune --omit=dev" drops tsx, drizzle-kit and typescript. The migration
scripts need drizzle-orm and better-sqlite3, both regular dependencies, so
they keep working. The dev target keeps devDependencies for compose.dev.yaml.

Both compose files get a mem_limit. The server stack also sets memswap_limit
to the same value, so a leak lets the OOM killer replace the container
instead of pushing the host into swap. The Express API is no longer
published on the host: it has no authentication and is reached through the
Next.js rewrite. Development ports are bound to the loopback interface.
Services drop all capabilities and run with no-new-privileges.

SIGTERM and SIGINT now close the HTTP server and the SQLite handle and exit,
instead of only logging the signal and waiting for the container's SIGKILL
ten seconds later. server.closeIdleConnections() releases the keep-alive
sockets the Next.js server holds, which would otherwise keep server.close()
pending for the whole grace period. The compose start commands exec into
node, because otherwise sh stays the foreground process and the signal never
reaches it.

Delete scripts/docker-start.sh rather than fixing its unsupervised
background API: both compose stacks set their own command, so the
single-container entrypoint has no caller left. Also remove drizzle.config.js
(a stray tsc artifact drizzle-kit never reads) and the pre-Next index.html
and styles.css.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

compose.dev.yaml

@@ -1,35 +1,47 @@
-# Development stack: source mounts, watching dev servers, hot reload.
+# Entwicklungs-Stack: Quellcode-Mounts, laufende Dev-Server, Hot Reload.
 #   docker compose -f compose.dev.yaml up --build
 #
-# The polling watchers below are needed for bind mounts on Windows and
-# macOS, where inotify events do not cross the VM boundary. They cost
-# continuous CPU, which is why the production stack in compose.yaml does
-# not run watchers at all.
+# Die Polling-Watcher unten sind fuer Bind-Mounts unter Windows und macOS
+# noetig, weil inotify-Events die VM-Grenze nicht passieren. Sie kosten
+# dauerhaft CPU; genau deshalb laufen im Server-Stack (compose.yaml) gar
+# keine Watcher.
+#
+# Der Projektname muss sich von dem des Server-Stacks unterscheiden, damit
+# "docker compose down" in diesem Verzeichnis niemals einen laufenden
+# Server-Stack abraeumen kann.
 name: leistungsbilanz-dev
 
-x-logging: &logging
-  driver: json-file
-  options:
-    max-size: "20m"
-    max-file: "10"
+x-service-defaults: &service-defaults
+  init: true
+  restart: unless-stopped
+  stop_grace_period: 20s
+  logging:
+    driver: json-file
+    options:
+      max-size: "20m"
+      max-file: "10"
 
 services:
   api:
+    <<: *service-defaults
     build:
       context: .
+      target: dev
     command:
       - sh
       - -c
-      - npm run db:migrate && npm run db:verify:circuit-schema && npm run dev:api
+      - npm run db:migrate && npm run db:verify:circuit-schema && exec npm run dev:api
     environment:
       PORT: "3000"
       CHOKIDAR_USEPOLLING: "true"
       LOG_LEVEL: "${LOG_LEVEL:-debug}"
-    init: true
-    restart: unless-stopped
-    logging: *logging
+    # Nur auf dem Loopback: nichts davon ist authentifiziert, die API darf
+    # nicht aus dem LAN erreichbar sein.
     ports:
-      - "3000:3000"
+      - "127.0.0.1:3000:3000"
+    # Kein memswap_limit wie im Server-Stack: beim Entwickeln ist Swap als
+    # Puffer erwuenscht, die Grenze soll nur den Host vor Dauerlast schuetzen.
+    mem_limit: 1g
     volumes:
       - ./src:/app/src
       - ./scripts:/app/scripts
@@ -41,15 +53,17 @@ services:
         - CMD
         - node
         - -e
-        - fetch('http://localhost:3000/health').then(response=>{if(!response.ok)process.exit(1)}).catch(()=>process.exit(1))
+        - fetch('http://127.0.0.1:3000/health').then(response=>{if(!response.ok)process.exit(1)}).catch(()=>process.exit(1))
       interval: 30s
       timeout: 3s
       retries: 5
       start_period: 20s
 
   web:
+    <<: *service-defaults
     build:
       context: .
+      target: dev
     command:
       - npm
       - run
@@ -62,14 +76,14 @@ services:
       WATCHPACK_POLLING: "true"
       NEXT_TELEMETRY_DISABLED: "1"
       LOG_LEVEL: "${LOG_LEVEL:-debug}"
-    init: true
-    restart: unless-stopped
-    logging: *logging
     depends_on:
       api:
         condition: service_healthy
     ports:
-      - "3001:3001"
+      - "127.0.0.1:3001:3001"
+    # "next dev" haelt Kompilierzustand und waechst im Dauerbetrieb stetig;
+    # die Grenze macht daraus einen Container-Neustart statt Host-Swapping.
+    mem_limit: 2g
     volumes:
       - ./src:/app/src
       - ./next.config.mjs:/app/next.config.mjs:ro
@@ -80,7 +94,7 @@ services:
         - CMD
         - node
         - -e
-        - fetch('http://localhost:3001/web-health').then(response=>{if(!response.ok)process.exit(1)}).catch(()=>process.exit(1))
+        - fetch('http://127.0.0.1:3001/web-health').then(response=>{if(!response.ok)process.exit(1)}).catch(()=>process.exit(1))
       interval: 30s
       timeout: 3s
       retries: 5