docker-compose.observability.yml - master - Al/TESamples

docker-compose.observability.yml
187 строк8 KB
Initial Commit
11 май 2026, 17:40
11 май 2026, 17:400a04577
# =============================================================
# Observability overlay (Phase 4)
#
# Использование:
#   docker compose -f docker-compose.yml -f docker-compose.observability.yml up -d
#   make up-obs
#
# Подключает:
#   - OTel Collector (OTLP gRPC :4317, HTTP :4318)
#   - Jaeger 1.6x (UI :16686, OTLP :4317 внутри сети)
#   - Prometheus (UI :9090, scrape Spring/Python/Keycloak)
#   - Grafana (UI :3000, datasources + dashboards заранее предзагружены)
#
# А также — добавляет к Java/Python сервисам OTLP-эндпоинт и включает
# Keycloak metrics на management-порту (KC_METRICS_ENABLED=true).
# =============================================================

services:
  # ---------- Включаем метрики Keycloak (FR-09.1) ----------
  # По умолчанию метрики Keycloak выключены. Здесь включаем на management-порту
  # :9000 (см. docker-compose.yml → KC_HTTP_MANAGEMENT_PORT). Prometheus
  # скрейпит /metrics на этом порту.
  keycloak:
    environment:
      KC_METRICS_ENABLED: "true"

  # ---------- OTel Collector ----------
  # Центральная точка приёма трейсов от Java/Python приложений. Batch + resource
  # processors применяются перед форвардом в Jaeger (см. otel-collector-config.yml).
  otel-collector:
    image: otel/opentelemetry-collector-contrib:0.116.1
    restart: unless-stopped
    command: ["--config=/etc/otel/config.yml"]
    volumes:
      - ./observability/otel-collector/otel-collector-config.yml:/etc/otel/config.yml:ro
    ports:
      - "4317:4317"   # OTLP gRPC (для отладки с хоста)
      - "4318:4318"   # OTLP HTTP
    healthcheck:
      # Health-extension коллектора на :13133 — простой 200 OK при готовности.
      test: ["CMD-SHELL", "wget -qO- http://localhost:13133/ || exit 1"]
      interval: 5s
      timeout: 3s
      retries: 10

  # ---------- Jaeger (all-in-one с OTLP receiver) ----------
  # all-in-one — collector+query+UI в одном процессе, без отдельной БД (in-memory).
  # Для dev достаточно. В production — раздельные jaeger-collector + storage.
  jaeger:
    image: jaegertracing/all-in-one:1.62.0
    restart: unless-stopped
    environment:
      # Включает OTLP-receiver на :4317/:4318 (отдельно от Jaeger-native ports).
      COLLECTOR_OTLP_ENABLED: "true"
    ports:
      - "${JAEGER_UI_PORT:-16686}:16686"

  # ---------- Prometheus ----------
  prometheus:
    image: prom/prometheus:v2.55.1
    restart: unless-stopped
    command:
      - "--config.file=/etc/prometheus/prometheus.yml"
      - "--storage.tsdb.path=/prometheus"
      # 24h в dev достаточно для разбора инцидента. В production — обычно ≥30d.
      - "--storage.tsdb.retention.time=24h"
    volumes:
      - ./observability/prometheus/prometheus.yml:/etc/prometheus/prometheus.yml:ro
      - ./observability/prometheus/alerts.yml:/etc/prometheus/alerts.yml:ro
      - prometheus_data:/prometheus
    ports:
      - "${PROMETHEUS_PORT:-9090}:9090"

  # ---------- Grafana ----------
  grafana:
    image: grafana/grafana:11.3.0
    restart: unless-stopped
    environment:
      # Учётка admin — dev-default. В production: ротация + Vault.
      GF_SECURITY_ADMIN_USER: ${GF_ADMIN_USER:-admin}
      GF_SECURITY_ADMIN_PASSWORD: ${GF_ADMIN_PASSWORD:-admin-dev}
      # Никаких self-signup'ов: пользователей заводит только admin.
      GF_USERS_ALLOW_SIGN_UP: "false"
      # Анонимный доступ только в роли Viewer — удобно для демо без логина.
      # В production: false.
      GF_AUTH_ANONYMOUS_ENABLED: "true"
      GF_AUTH_ANONYMOUS_ORG_ROLE: "Viewer"
    volumes:
      # provisioning: автозагрузка datasources/dashboards-конфигов из git.
      - ./observability/grafana/provisioning:/etc/grafana/provisioning:ro
      # Сами JSON-дашборды (путь должен совпадать с dashboards.yml options.path).
      - ./observability/grafana/dashboards:/var/lib/grafana/dashboards:ro
      - grafana_data:/var/lib/grafana
    ports:
      - "${GRAFANA_PORT:-3000}:3000"
    depends_on:
      prometheus:
        condition: service_started

  # ---------- Дополняем Java сервисы OTLP-конфигурацией ----------
  # MANAGEMENT_OTLP_TRACING_ENDPOINT через relaxed-binding выставляет
  # management.otlp.tracing.endpoint в Spring (см. application.yml).
  # Sampling=1.0 в dev: все трейсы в Jaeger. В production — 0.01–0.1.
  bff-agent:
    environment:
      MANAGEMENT_OTLP_TRACING_ENDPOINT: "http://otel-collector:4318/v1/traces"
      MANAGEMENT_TRACING_SAMPLING_PROBABILITY: "1.0"
      OTEL_SERVICE_NAME: "bff-agent"
    depends_on:
      otel-collector:
        condition: service_started

  service-a:
    environment:
      MANAGEMENT_OTLP_TRACING_ENDPOINT: "http://otel-collector:4318/v1/traces"
      MANAGEMENT_TRACING_SAMPLING_PROBABILITY: "1.0"
      OTEL_SERVICE_NAME: "service-a"
    depends_on:
      otel-collector:
        condition: service_started

  service-b:
    environment:
      MANAGEMENT_OTLP_TRACING_ENDPOINT: "http://otel-collector:4318/v1/traces"
      MANAGEMENT_TRACING_SAMPLING_PROBABILITY: "1.0"
      OTEL_SERVICE_NAME: "service-b"

  service-c:
    environment:
      MANAGEMENT_OTLP_TRACING_ENDPOINT: "http://otel-collector:4318/v1/traces"
      MANAGEMENT_TRACING_SAMPLING_PROBABILITY: "1.0"
      OTEL_SERVICE_NAME: "service-c"

  # ---------- Дополняем Python сервисы OTel auto-instrumentation ----------
  # Наличие OTEL_EXPORTER_OTLP_ENDPOINT включает запуск под
  # `opentelemetry-instrument` (см. apps/python/Dockerfile ENTRYPOINT).
  # METRICS_EXPORTER=none и LOGS_EXPORTER=none, потому что метрики собираем
  # отдельно (Prometheus scrape /metrics), а логи — через docker logs.
  bff-agent-py:
    environment:
      OTEL_SERVICE_NAME: "bff-agent-py"
      OTEL_EXPORTER_OTLP_ENDPOINT: "http://otel-collector:4317"
      OTEL_EXPORTER_OTLP_PROTOCOL: "grpc"
      OTEL_TRACES_EXPORTER: "otlp"
      OTEL_METRICS_EXPORTER: "none"
      OTEL_LOGS_EXPORTER: "none"
      OTEL_INSTRUMENTATION_ENABLED: "true"
    depends_on:
      otel-collector:
        condition: service_started

  service-a-py:
    environment:
      OTEL_SERVICE_NAME: "service-a-py"
      OTEL_EXPORTER_OTLP_ENDPOINT: "http://otel-collector:4317"
      OTEL_EXPORTER_OTLP_PROTOCOL: "grpc"
      OTEL_TRACES_EXPORTER: "otlp"
      OTEL_METRICS_EXPORTER: "none"
      OTEL_LOGS_EXPORTER: "none"
      OTEL_INSTRUMENTATION_ENABLED: "true"
    depends_on:
      otel-collector:
        condition: service_started

  service-b-py:
    environment:
      OTEL_SERVICE_NAME: "service-b-py"
      OTEL_EXPORTER_OTLP_ENDPOINT: "http://otel-collector:4317"
      OTEL_EXPORTER_OTLP_PROTOCOL: "grpc"
      OTEL_TRACES_EXPORTER: "otlp"
      OTEL_METRICS_EXPORTER: "none"
      OTEL_LOGS_EXPORTER: "none"
      OTEL_INSTRUMENTATION_ENABLED: "true"

  service-c-py:
    environment:
      OTEL_SERVICE_NAME: "service-c-py"
      OTEL_EXPORTER_OTLP_ENDPOINT: "http://otel-collector:4317"
      OTEL_EXPORTER_OTLP_PROTOCOL: "grpc"
      OTEL_TRACES_EXPORTER: "otlp"
      OTEL_METRICS_EXPORTER: "none"
      OTEL_LOGS_EXPORTER: "none"
      OTEL_INSTRUMENTATION_ENABLED: "true"

volumes:
  prometheus_data:
  grafana_data: