docker-compose.observability.yml

# Observability stack for Nexus (Issue #761)
#
# Deploys Grafana, Prometheus, Loki, Tempo, and Grafana Alloy alongside
# the Nexus server.  All services share the `nexus-network` from the
# main docker-compose.yml.
#
# Usage:
#   docker compose -f docker-compose.yml -f docker-compose.observability.yml up -d
#
# Or standalone (Nexus must already be running on nexus-network):
#   docker compose -f docker-compose.observability.yml up -d

version: "3.8"

services:
  # =========================================================================
  # Nexus Server (extends base compose with observability env vars)
  # =========================================================================
  nexus-server:
    environment:
      - OTEL_ENABLED=true
      - OTEL_EXPORTER_OTLP_ENDPOINT=http://tempo:4317
      - OTEL_SERVICE_NAME=nexus
      - NEXUS_ENV=prod
      - PYROSCOPE_ENABLED=true
      - PYROSCOPE_SERVER_ADDRESS=http://pyroscope:4040
    networks:
      - nexus-network

  # =========================================================================
  # Prometheus — Metric collection and storage
  # =========================================================================
  prometheus:
    image: prom/prometheus:latest
    container_name: nexus-prometheus
    restart: unless-stopped
    ports:
      - "9090:9090"
    volumes:
      - ./observability/prometheus/prometheus.yml:/etc/prometheus/prometheus.yml:ro
      - prometheus-data:/prometheus
    command:
      - "--config.file=/etc/prometheus/prometheus.yml"
      - "--storage.tsdb.retention.time=7d"
      - "--web.enable-remote-write-receiver"
    deploy:
      resources:
        limits:
          memory: 512M
    networks:
      - nexus-network
    profiles:
      - observability

  # =========================================================================
  # Loki — Log aggregation
  # =========================================================================
  loki:
    image: grafana/loki:3.0.0
    container_name: nexus-loki
    restart: unless-stopped
    ports:
      - "3100:3100"
    volumes:
      - ./observability/loki/loki.yml:/etc/loki/config.yaml:ro
      - loki-data:/loki
    command:
      - "-config.file=/etc/loki/config.yaml"
    deploy:
      resources:
        limits:
          memory: 512M
    networks:
      - nexus-network
    profiles:
      - observability

  # =========================================================================
  # Tempo — Distributed tracing
  # =========================================================================
  tempo:
    image: grafana/tempo:latest
    container_name: nexus-tempo
    restart: unless-stopped
    ports:
      - "3200:3200"   # Tempo HTTP API
      - "4317:4317"   # OTLP gRPC receiver
      - "4318:4318"   # OTLP HTTP receiver
    volumes:
      - ./observability/tempo/tempo.yml:/etc/tempo/config.yaml:ro
      - tempo-data:/tmp/tempo
    command:
      - "-config.file=/etc/tempo/config.yaml"
    deploy:
      resources:
        limits:
          memory: 512M
    networks:
      - nexus-network
    profiles:
      - observability

  # =========================================================================
  # Grafana — Visualization and dashboards
  # =========================================================================
  grafana:
    image: grafana/grafana:latest
    container_name: nexus-grafana
    restart: unless-stopped
    ports:
      - "3000:3000"
    environment:
      - GF_SECURITY_ADMIN_USER=admin
      - GF_SECURITY_ADMIN_PASSWORD=nexus
      - GF_USERS_ALLOW_SIGN_UP=false
      - GF_AUTH_ANONYMOUS_ENABLED=true
      - GF_AUTH_ANONYMOUS_ORG_ROLE=Viewer
    volumes:
      - ./observability/grafana/provisioning:/etc/grafana/provisioning:ro
      - grafana-data:/var/lib/grafana
    deploy:
      resources:
        limits:
          memory: 256M
    networks:
      - nexus-network
    depends_on:
      - prometheus
      - loki
      - tempo
      - pyroscope
    profiles:
      - observability

  # =========================================================================
  # Grafana Alloy — Log collection agent (Promtail successor)
  # =========================================================================
  alloy:
    image: grafana/alloy:latest
    container_name: nexus-alloy
    restart: unless-stopped
    ports:
      - "12345:12345"  # Alloy UI
    volumes:
      - ./observability/alloy/config.alloy:/etc/alloy/config.alloy:ro
      - /var/run/docker.sock:/var/run/docker.sock:ro
    command:
      - "run"
      - "/etc/alloy/config.alloy"
      - "--server.http.listen-addr=0.0.0.0:12345"
    deploy:
      resources:
        limits:
          memory: 128M
    networks:
      - nexus-network
    depends_on:
      - loki
    profiles:
      - observability

  # =========================================================================
  # postgres_exporter — PostgreSQL metrics for Prometheus (Issue #762)
  # =========================================================================
  postgres-exporter:
    image: quay.io/prometheuscommunity/postgres-exporter:latest
    container_name: nexus-postgres-exporter
    restart: unless-stopped
    environment:
      DATA_SOURCE_NAME: "postgresql://nexus_monitor:${POSTGRES_EXPORTER_PASSWORD:-nexus_monitor}@postgres:5432/${POSTGRES_DB:-nexus}?sslmode=disable"
    command:
      - "--collector.stat_statements"
      - "--collector.stat_activity"
      - "--collector.stat_bgwriter"
      - "--collector.stat_database"
      - "--collector.stat_user_tables"
      - "--collector.locks"
      - "--collector.wal"
    ports:
      - "9187:9187"
    deploy:
      resources:
        limits:
          memory: 128M
    networks:
      - nexus-network
    profiles:
      - observability

  # =========================================================================
  # Pyroscope — Continuous profiling (Issue #763)
  # =========================================================================
  pyroscope:
    image: grafana/pyroscope:1.16.1
    container_name: nexus-pyroscope
    restart: unless-stopped
    ports:
      - "4040:4040"
    volumes:
      - ./observability/pyroscope/pyroscope.yml:/etc/pyroscope/config.yaml:ro
      - pyroscope-data:/data
    command:
      - "-config.file=/etc/pyroscope/config.yaml"
    deploy:
      resources:
        limits:
          memory: 512M
    networks:
      - nexus-network
    profiles:
      - observability

volumes:
  prometheus-data:
  loki-data:
  tempo-data:
  pyroscope-data:
  grafana-data:

networks:
  nexus-network:
    external: true