feat: initial commit with vLLM production stack and benchmark tools

2026-09-09 18:22:40 +02:00
commit 20ae3cfb48
10 changed files with 331 additions and 0 deletions
@@ -0,0 +1,101 @@
+version: '3.8'
+
+services:
+  # 1. MOTOR DE INFERENCIA DE PRODUCCIÓN (vLLM)
+  vllm:
+    image: vllm/vllm-openai:latest
+    container_name: vllm-engine
+    restart: unless-stopped
+    environment:
+      - HUGGING_FACE_HUB_TOKEN=${HF_TOKEN}
+      - VLLM_LOGGING_LEVEL=INFO
+    volumes:
+      - ./models_cache:/root/.cache/huggingface:Z # :Z para compatibilidad SELinux en RHEL/Fedora
+    ports:
+      - "8000:8000"
+    command: >
+      --model ${MODEL_ID}
+      --quantization awq
+      --dtype half
+      --gpu-memory-utilization ${GPU_MEMORY_UTILIZATION}
+      --max-model-len ${MAX_MODEL_LEN}
+      --max-num-seqs ${MAX_NUM_SEQS}
+      --block-size 16
+      --port 8000
+    deploy:
+      resources:
+        reservations:
+          devices:
+            - driver: nvidia
+              count: all
+              capabilities: [gpu]
+    networks:
+      - vllm-net
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
+      interval: 10s
+      timeout: 5s
+      retries: 10
+      start_period: 60s
+
+  # 2. PROXY DE RED, RESILIENCIA Y CIRCUIT BREAKER (LiteLLM)
+  litellm:
+    image: ghcr.io/berriai/litellm:main-latest
+    container_name: litellm-gateway
+    restart: unless-stopped
+    volumes:
+      - ./litellm/config.yaml:/app/config.yaml:ro,Z
+    ports:
+      - "4000:4000"
+    command: ["--config", "/app/config.yaml", "--port", "4000"]
+    depends_on:
+      vllm:
+        condition: service_healthy
+    networks:
+      - vllm-net
+
+  # 3. OBSERVABILIDAD: RECOLECTOR DE TELEMETRÍA (Prometheus)
+  prometheus:
+    image: prom/prometheus:latest
+    container_name: prometheus-telemetry
+    restart: unless-stopped
+    volumes:
+      - ./prometheus/prometheus.yml:/etc/prometheus/prometheus.yml:ro,Z
+      - prometheus-data:/prometheus:Z
+    ports:
+      - "9090:9090"
+    command:
+      - '--config.file=/etc/prometheus/prometheus.yml'
+      - '--storage.tsdb.path=/prometheus'
+      - '--web.console.libraries=/usr/share/prometheus/console_libraries'
+      - '--web.console.templates=/usr/share/prometheus/consoles'
+    networks:
+      - vllm-net
+
+  # 4. OBSERVABILIDAD: VISUALIZACIÓN EN TIEMPO REAL (Grafana)
+  grafana:
+    image: grafana/grafana:latest
+    container_name: grafana-dashboard
+    restart: unless-stopped
+    environment:
+      - GF_SECURITY_ADMIN_USER=admin
+      - GF_SECURITY_ADMIN_PASSWORD=admin
+      - GF_USERS_ALLOW_SIGN_UP=false
+    volumes:
+      - ./grafana/provisioning:/etc/grafana/provisioning:ro,Z
+      - grafana-data:/var/lib/grafana:Z
+    ports:
+      - "3000:3000"
+    depends_on:
+      - prometheus
+    networks:
+      - vllm-net
+
+networks:
+  vllm-net:
+    driver: bridge
+
+volumes:
+  prometheus-data:
+  grafana-data:
+