Initial commit

infra/base/monitoring.yaml

@@ -0,0 +1,126 @@
+# FILE: infra/compose/production/monitoring.yaml
+# Production Monitoring Stack for AI Tax Agent
+# Deploy to: /opt/compose/ai-tax-agent/monitoring.yaml
+
+networks:
+  frontend:
+    external: true
+    name: frontend
+  backend:
+    external: true
+    name: backend
+
+volumes:
+  prometheus_data:
+  grafana_data:
+  loki_data:
+
+services:
+  # Metrics Collection
+  prometheus:
+    image: prom/prometheus:v2.48.1
+    container_name: prometheus
+    restart: unless-stopped
+    networks:
+      - backend
+      - frontend
+    volumes:
+      - prometheus_data:/prometheus
+      - ./prometheus/prometheus.yml:/etc/prometheus/prometheus.yml:ro
+    command:
+      - "--config.file=/etc/prometheus/prometheus.yml"
+      - "--storage.tsdb.path=/prometheus"
+      - "--web.console.libraries=/etc/prometheus/console_libraries"
+      - "--web.console.templates=/etc/prometheus/consoles"
+      - "--storage.tsdb.retention.time=30d"
+      - "--web.enable-lifecycle"
+    labels:
+      - "traefik.enable=true"
+      - "traefik.http.routers.prometheus.rule=Host(`prometheus.${DOMAIN}`)"
+      - "traefik.http.routers.prometheus.entrypoints=websecure"
+      - "traefik.http.routers.prometheus.tls=true"
+      - "traefik.http.routers.prometheus.tls.certresolver=godaddy"
+      - "traefik.http.routers.prometheus.middlewares=authentik-forwardauth@file"
+      - "traefik.http.services.prometheus.loadbalancer.server.port=9090"
+
+  # Visualization & Dashboards
+  grafana:
+    image: grafana/grafana:10.2.3
+    container_name: grafana
+    restart: unless-stopped
+    networks:
+      - backend
+      - frontend
+    volumes:
+      - grafana_data:/var/lib/grafana
+      - ./grafana/provisioning:/etc/grafana/provisioning:ro
+      - ./grafana/dashboards:/var/lib/grafana/dashboards:ro
+    environment:
+      GF_SECURITY_ADMIN_PASSWORD: ${GRAFANA_PASSWORD}
+      GF_USERS_ALLOW_SIGN_UP: false
+      GF_USERS_AUTO_ASSIGN_ORG: true
+      GF_USERS_AUTO_ASSIGN_ORG_ROLE: Viewer
+      GF_AUTH_GENERIC_OAUTH_ENABLED: true
+      GF_AUTH_GENERIC_OAUTH_NAME: Authentik
+      GF_AUTH_GENERIC_OAUTH_CLIENT_ID: ${GRAFANA_OAUTH_CLIENT_ID}
+      GF_AUTH_GENERIC_OAUTH_CLIENT_SECRET: ${GRAFANA_OAUTH_CLIENT_SECRET}
+      GF_AUTH_GENERIC_OAUTH_SCOPES: openid profile email groups
+      GF_AUTH_GENERIC_OAUTH_AUTH_URL: https://authentik.${DOMAIN}/application/o/authorize/
+      GF_AUTH_GENERIC_OAUTH_TOKEN_URL: https://authentik.${DOMAIN}/application/o/token/
+      GF_AUTH_GENERIC_OAUTH_API_URL: https://authentik.${DOMAIN}/application/o/userinfo/
+      GF_AUTH_GENERIC_OAUTH_AUTO_LOGIN: false
+      GF_AUTH_GENERIC_OAUTH_ALLOW_SIGN_UP: true
+      GF_AUTH_GENERIC_OAUTH_ROLE_ATTRIBUTE_PATH: role
+      GF_AUTH_GENERIC_OAUTH_ROLE_ATTRIBUTE_STRICT: false
+      GF_AUTH_GENERIC_OAUTH_GROUPS_ATTRIBUTE_PATH: groups
+      GF_AUTH_OAUTH_AUTO_LOGIN: false
+      GF_AUTH_DISABLE_LOGIN_FORM: false
+      GF_SERVER_ROOT_URL: https://grafana.${DOMAIN}
+      GF_SERVER_SERVE_FROM_SUB_PATH: false
+      GF_SECURITY_COOKIE_SECURE: true
+      GF_SECURITY_COOKIE_SAMESITE: lax
+      GF_AUTH_GENERIC_OAUTH_USE_PKCE: true
+    labels:
+      - "traefik.enable=true"
+      - "traefik.http.routers.grafana.rule=Host(`grafana.${DOMAIN}`)"
+      - "traefik.http.routers.grafana.entrypoints=websecure"
+      - "traefik.http.routers.grafana.tls=true"
+      - "traefik.http.routers.grafana.tls.certresolver=godaddy"
+      - "traefik.http.services.grafana.loadbalancer.server.port=3000"
+
+  # Log Aggregation
+  loki:
+    image: grafana/loki:2.9.4
+    container_name: loki
+    restart: unless-stopped
+    networks:
+      - backend
+      - frontend
+    volumes:
+      - loki_data:/loki
+      - ./loki/loki.yml:/etc/loki/local-config.yaml:ro
+    command: -config.file=/etc/loki/local-config.yaml
+    labels:
+      - "traefik.enable=true"
+      - "traefik.http.routers.loki.rule=Host(`loki.${DOMAIN}`)"
+      - "traefik.http.routers.loki.entrypoints=websecure"
+      - "traefik.http.routers.loki.tls=true"
+      - "traefik.http.routers.loki.tls.certresolver=godaddy"
+      - "traefik.http.routers.loki.middlewares=authentik-forwardauth@file"
+      - "traefik.http.services.loki.loadbalancer.server.port=3100"
+
+  # Log Shipper (for Docker containers)
+  promtail:
+    image: grafana/promtail:2.9.4
+    container_name: promtail
+    restart: unless-stopped
+    networks:
+      - backend
+    volumes:
+      - /var/log:/var/log:ro
+      - /var/lib/docker/containers:/var/lib/docker/containers:ro
+      - ./loki/promtail-config.yml:/etc/promtail/config.yml:ro
+    command: -config.file=/etc/promtail/config.yml
+    depends_on:
+      - loki
+