obmp-docker/docker-compose.yml

---
name: obmp
volumes:
  data-volume:
    driver_opts:
      type: none
      device: ${OBMP_DATA_ROOT}/postgres/data
      o: bind
  ts-volume:
    driver_opts:
      type: none
      device: ${OBMP_DATA_ROOT}/postgres/ts
      o: bind

services:

  zookeeper:
    restart: unless-stopped
    container_name: obmp-zookeeper
    healthcheck:
      test: ["CMD-SHELL", "bash -c 'echo > /dev/tcp/localhost/2181'"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 60s
    image: confluentinc/cp-zookeeper:7.1.1
    mem_limit: 1g
    volumes:
      - ${OBMP_DATA_ROOT}/zk-data:/var/lib/zookeeper/data
      - ${OBMP_DATA_ROOT}/zk-log:/var/lib/zookeeper/log
    environment:
      ZOOKEEPER_CLIENT_PORT: 2181
      ZOOKEEPER_TICK_TIME: 2000

  kafka:
    restart: unless-stopped
    container_name: obmp-kafka
    healthcheck:
      test: ["CMD-SHELL", "bash -c 'echo > /dev/tcp/localhost/9092'"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 90s
    image: confluentinc/cp-kafka:7.1.1
    # Raise KAFKA_MEM_LIMIT for production (full-table initial dumps are bursty).
    mem_limit: ${KAFKA_MEM_LIMIT:-4g}

    # Change the mount point to where you want to store Kafka data.
    #   Normally 80GB or more
    volumes:
      - ${OBMP_DATA_ROOT}/kafka-data:/var/lib/kafka/data
    depends_on:
      - zookeeper
    ports:
      - "9092:9092"
    environment:
      KAFKA_BROKER_ID: 1
      KAFKA_ZOOKEEPER_CONNECT: obmp-zookeeper:2181

      # Change/add listeners based on your FQDN that the host and other containers can access.  You can use
      #    an IP address as well. By default, only within the compose/containers can Kafka be accesssed
      #    using port 29092.   Outside access can be enabled, but you should use an FQDN listener.
      KAFKA_ADVERTISED_LISTENERS: PLAINTEXT://obmp-kafka:29092,PLAINTEXT_HOST://${HOST_IP:-10.40.40.202}:9092
      KAFKA_LISTENER_SECURITY_PROTOCOL_MAP: PLAINTEXT:PLAINTEXT,PLAINTEXT_HOST:PLAINTEXT
      KAFKA_INTER_BROKER_LISTENER_NAME: PLAINTEXT
      KAFKA_OFFSETS_TOPIC_REPLICATION_FACTOR: 1
      KAFKA_NUM_PARTITIONS: 8
      KAFKA_LOG_RETENTION_MINUTES: 720
      KAFKA_LOG_ROLL_MS: 3600000
      KAFKA_LOG_SEGMENT_BYTES: 1073741824
      KAFKA_MESSAGE_MAX_BYTES: 100000000
      KAFKA_LOG_CLEANER_THREADS: 2

  # TODO: Move comment to documentation
  # You can use SWAG and Let's Encrypt to front the Grafana HTTP port 3000 instance. Below is an example
  # Create the SWAG direcgtory using "sudo mkdir -m 777 ${OBMP_DATA_ROOT}/swag"
#  swag:
#    image: linuxserver/letsencrypt:version-1.11.0
#    container_name: obmp-swag
#    cap_add:
#      - NET_ADMIN
#    environment:
#      - PUID=1000
#      - PGID=1000
#      - TZ=UTC
#      - URL=<domain of grafana host, this should be the domain only, not the hostname>
#      - SUBDOMAINS=<hostname, without the domain>,
#      - VALIDATION=http
#      - EMAIL=<your email>
#      - ONLY_SUBDOMAINS=true
#      - STAGING=false
#    volumes:
#      - /var/openbmp/swag:/config
#    ports:
#      - 443:443
#      - 80:80 #optional
#    restart: unless-stopped

  grafana:
    restart: unless-stopped
    container_name: obmp-grafana
    healthcheck:
      test: ["CMD-SHELL", "wget -q --spider http://localhost:3000/api/health || exit 1"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 40s
    image: grafana/grafana:9.1.7
    mem_limit: 1g
    ports:
      - "3000:3000"
    volumes:
      - ${OBMP_DATA_ROOT}/grafana:/var/lib/grafana
      - ${OBMP_DATA_ROOT}/grafana/provisioning:/etc/grafana/provisioning/
    environment:
      - GF_SECURITY_ADMIN_PASSWORD=openbmp
      - GF_AUTH_ANONYMOUS_ENABLED=false
      - GF_SERVER_ROOT_URL=https://${OBMP_DOMAIN:-bmp.apodacalab.com}/grafana/
      - GF_SERVER_SERVE_FROM_SUB_PATH=true
      - GF_AUTH_PROXY_ENABLED=true
      - GF_AUTH_PROXY_HEADER_NAME=Remote-User
      - GF_AUTH_PROXY_HEADER_PROPERTY=username
      - GF_AUTH_PROXY_AUTO_SIGN_UP=true
      - GF_USERS_HOME_PAGE=d/obmp-home/obmp-home
      - GF_INSTALL_PLUGINS=agenty-flowcharting-panel,grafana-piechart-panel,grafana-worldmap-panel,grafana-simple-json-datasource,vonage-status-panel

      # TODO: Move comment block to documentation
      # The below is an example of how to enable OAuth
      #- GF_AUTH_GENERIC_OAUTH_NAME=MyCompany SSO
      #- GF_AUTH_GENERIC_OAUTH_ICON=signin
      #- GF_AUTH_GENERIC_OAUTH_ENABLED=true
      #- GF_AUTH_GENERIC_OAUTH_CLIENT_ID=<the client ID>
      #- GF_AUTH_GENERIC_OAUTH_CLIENT_SECRET=<client secret>
      #- GF_AUTH_GENERIC_OAUTH_ALLOW_SIGN_UP=true
      #- GF_AUTH_GENERIC_OAUTH_AUTH_URL=<auth url, such as https://domain/as/authorization.oauth2>
      #- GF_AUTH_GENERIC_OAUTH_TOKEN_URL=<token url, such as https://domain/as/token.oauth2
      #- GF_AUTH_GENERIC_OAUTH_API_URL=<userinfo url, such as https://domain/idp/userinfo.openid
      #- GF_AUTH_GENERIC_OAUTH_SCOPES=email openid profile
      #- GF_AUTH_GENERIC_OAUTH_SEND_CLIENT_CREDENTIALS_VIA_POST=true
      #- GF_SERVER_ENFORCE_DOMAIN=true
      #- GF_SERVER_PROTOCOL=http
      #- GF_SERVER_HTTP_PORT=3000
      #- GF_SERVER_DOMAIN=<fqdn of grafana server>
      #- GF_SERVER_ROOT_URL=%(protocol)s://%(domain)s:%(http_port)s/

  psql:
    restart: unless-stopped
    container_name: obmp-psql
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U openbmp -d openbmp"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 60s
    image: openbmp/postgres:2.2.1
    # Raise PSQL_MEM_LIMIT for production (see docs/production-sizing.md).
    mem_limit: ${PSQL_MEM_LIMIT:-6g}
    privileged: true
    shm_size: 1536m
    sysctls:
      - net.ipv4.tcp_keepalive_intvl=30
      - net.ipv4.tcp_keepalive_probes=5
      - net.ipv4.tcp_keepalive_time=180
    ports:
      - "5432:5432"
    volumes:
      - data-volume:/var/lib/postgresql/data
      - ts-volume:/var/lib/postgresql/ts
    # alter_job max_runtime in _timescaledb_config.bgw_job ( https://docs.timescale.com/latest/api#alter_job )
    command: >
      -c max_wal_size=10GB
    environment:
      - POSTGRES_PASSWORD=openbmp
      - POSTGRES_USER=openbmp
      - POSTGRES_DB=openbmp

  collector:
    restart: unless-stopped
    container_name: obmp-collector
    healthcheck:
      test: ["CMD-SHELL", "bash -c 'echo > /dev/tcp/localhost/5000'"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 40s
    image: openbmp/collector:2.2.3
    mem_limit: 2g
    sysctls:
      - net.ipv4.tcp_keepalive_intvl=30
      - net.ipv4.tcp_keepalive_probes=5
      - net.ipv4.tcp_keepalive_time=180
    ports:
      - "5000:5000"
    volumes:
      - ${OBMP_DATA_ROOT}/config:/config
    environment:
      - KAFKA_FQDN=obmp-kafka:29092

  psql-app:
    restart: unless-stopped
    container_name: obmp-psql-app
    # No healthcheck — the consumer exposes no health port; Docker's
    # restart-on-exit covers process death.
    image: openbmp/psql-app:2.2.2
    # mem_limit must exceed the MEM (JVM heap) env below. Raise both for
    # production — see docs/production-sizing.md.
    mem_limit: ${PSQL_APP_MEM_LIMIT:-4g}
    sysctls:
      - net.ipv4.tcp_keepalive_intvl=30
      - net.ipv4.tcp_keepalive_probes=5
      - net.ipv4.tcp_keepalive_time=180
    ports:
      - "9005:9005"

    volumes:
      - ${OBMP_DATA_ROOT}/config:/config
    environment:
      - MEM=3                                           # Set memory to at least 2GB but ideally 4GB
      - KAFKA_FQDN=obmp-kafka:29092
      - RPKI_URL=https://rpki.cloudflare.com/rpki.json  # define the URL to retrieve json endoed RPKI data
      - RPKI_PASS=None
      - RPKI_USER=None
      - ENABLE_RPKI=1                                   # 1 enables, 0 disables RPKI sync
      - ENABLE_IRR=1                                    # 1 enables, 0 disables IRR sync
      - ENABLE_DBIP=1                                   # 1 enables, 0 disables DBIP import
      - POSTGRES_REPORT_WINDOW='8 minute'               # default POSTGRESS window to select when building
                                                        # summary tables. For deployments that absorb large
                                                        # bursts increase the value, ex 60 minute
      - POSTGRES_PASSWORD=openbmp
      - POSTGRES_USER=openbmp
      - POSTGRES_DB=openbmp
      - POSTGRES_HOST=obmp-psql
      - POSTGRES_PORT=5432
      - POSTGRES_DROP_peer_event_log='1 year'
      - POSTGRES_DROP_stat_reports='4 weeks'
      - POSTGRES_DROP_ip_rib_log='4 weeks'
      - POSTGRES_DROP_alerts='4 weeks'
      - POSTGRES_DROP_ls_nodes_log='4 months'
      - POSTGRES_DROP_ls_links_log='4 months'
      - POSTGRES_DROP_ls_prefixes_log='4 months'
      - POSTGRES_DROP_stats_chg_byprefix='4 weeks'
      - POSTGRES_DROP_stats_chg_byasn='4 weeks'
      - POSTGRES_DROP_stats_chg_bypeer='4 weeks'
      - POSTGRES_DROP_stats_ip_origins='4 weeks'
      - POSTGRES_DROP_stats_peer_rib='4 weeks'
      - POSTGRES_DROP_stats_peer_update_counts='4 weeks'

  exabgp:
    restart: unless-stopped
    container_name: obmp-exabgp
    healthcheck:
      test: ["CMD-SHELL", "bash -c 'echo > /dev/tcp/localhost/5050'"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 40s
    profiles: ["test"]
    # The full-table feature generates up to 900K route objects in memory;
    # 512m OOM-killed it. Raise EXABGP_MEM_LIMIT in .env for larger tables.
    mem_limit: ${EXABGP_MEM_LIMIT:-6g}
    build:
      context: ./exabgp
      dockerfile: Dockerfile
    # Host networking so ExaBGP can reach CML routers directly on port 179
    network_mode: host
    environment:
      # IP on the host that CML routers reach (BGP peering source)
      - EXABGP_LOCAL_IP=${HOST_IP:-10.40.40.202}
      # ExaBGP presents as AS 65100 (eBGP peer to the lab route reflectors)
      - EXABGP_LOCAL_AS=${EXABGP_LOCAL_AS:-65100}
      # Peer list — ";"-separated entries of "ip:peer_as:description".
      # Default covers both labs: AS 65020 (ESXi) and AS 65021 (Proxmox).
      - EXABGP_PEERS=${EXABGP_PEERS:-10.100.0.100:65020:CML-R9K-CORE-01;10.100.0.200:65020:CML-R9K-CORE-02;10.100.1.100:65021:PROX-R9K-CORE-01;10.100.1.200:65021:PROX-R9K-CORE-02}
      # Flask API port (also on host network)
      - EXABGP_API_PORT=${EXABGP_API_PORT:-5050}
    volumes:
      # Mount scenarios dir so you can edit/add scenarios without rebuilding
      - ./exabgp/scenarios:/exabgp/scenarios
    # No ports: block needed — network_mode: host exposes directly

  exabgp-ui:
    restart: unless-stopped
    container_name: obmp-exabgp-ui
    healthcheck:
      test: ["CMD-SHELL", "wget -q --spider http://localhost:5001/ || exit 1"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 30s
    profiles: ["test"]
    mem_limit: 256m
    build:
      context: ./exabgp-ui
      dockerfile: Dockerfile
    # Host networking so NGINX can proxy /api to ExaBGP Flask on localhost:5050
    network_mode: host
    # Serves on port 5001 (host network, defined in nginx.conf)

  # --- Phase 4: gNMI Streaming Telemetry ---

  influxdb:
    restart: unless-stopped
    container_name: obmp-influxdb
    healthcheck:
      test: ["CMD-SHELL", "curl -fsS http://localhost:8086/health || exit 1"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 40s
    profiles: ["test"]
    image: influxdb:2.7
    mem_limit: 2g
    ports:
      - "8086:8086"
    volumes:
      - ${OBMP_DATA_ROOT}/influxdb:/var/lib/influxdb2
    environment:
      - DOCKER_INFLUXDB_INIT_MODE=setup
      - DOCKER_INFLUXDB_INIT_USERNAME=openbmp
      - DOCKER_INFLUXDB_INIT_PASSWORD=openbmp123
      - DOCKER_INFLUXDB_INIT_ORG=openbmp
      - DOCKER_INFLUXDB_INIT_BUCKET=telemetry
      - DOCKER_INFLUXDB_INIT_ADMIN_TOKEN=openbmp-telemetry-token
      - DOCKER_INFLUXDB_INIT_RETENTION=30d

  telegraf:
    restart: unless-stopped
    container_name: obmp-telegraf
    profiles: ["test"]
    mem_limit: 512m
    build:
      context: ./telegraf
      dockerfile: Dockerfile
    network_mode: host
    # Run telegraf as root and override the image entrypoint (which otherwise
    # drops back to the telegraf user) so [[inputs.docker]] can read the
    # Docker daemon socket for container resource metrics.
    user: root
    entrypoint: ["telegraf"]
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
      # Host root, read-only — lets [[inputs.disk]] report the real host
      # filesystems (Postgres/Kafka/InfluxDB data) instead of the container's.
      - /:/hostfs:ro
    depends_on:
      - influxdb
    environment:
      - INFLUXDB_TOKEN=openbmp-telemetry-token
      # Point gopsutil-based inputs (disk) at the host filesystem mount above.
      - HOST_MOUNT_PREFIX=/hostfs
      - HOST_PROC=/hostfs/proc
      - HOST_SYS=/hostfs/sys
      - HOST_ETC=/hostfs/etc
      # PostgreSQL credentials for [[inputs.postgresql_extensible]] (DB size).
      - POSTGRES_PASSWORD=${POSTGRES_PASSWORD:-openbmp}
      # gNMI fleet — quoted, comma-separated host:port list. Default = the two
      # ESXi CORE routers; extend via GNMI_ADDRESSES in .env for more routers.
      - 'GNMI_ADDRESSES=${GNMI_ADDRESSES:-"10.100.0.100:57400", "10.100.0.200:57400"}'
      - GNMI_USERNAME=${GNMI_USERNAME:-webui}
      - GNMI_PASSWORD=${GNMI_PASSWORD:-cisco}

  # --- Phase 4: Traffic Generator ---

  traffic-gen:
    restart: unless-stopped
    container_name: obmp-traffic-gen
    healthcheck:
      test: ["CMD-SHELL", "bash -c 'echo > /dev/tcp/localhost/5051'"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 30s
    profiles: ["test"]
    mem_limit: 1g
    build:
      context: ./traffic-gen
      dockerfile: Dockerfile
    network_mode: host
    cap_add:
      - NET_RAW
      - NET_ADMIN
    environment:
      - TRAFFIC_GEN_PORT=5051
      - TRAFFIC_GEN_MODE=sender
      - RESPONDER_URL=http://172.30.0.10:5053

  traffic-gen-ui:
    restart: unless-stopped
    container_name: obmp-traffic-gen-ui
    healthcheck:
      test: ["CMD-SHELL", "wget -q --spider http://localhost:5002/ || exit 1"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 30s
    profiles: ["test"]
    mem_limit: 256m
    build:
      context: ./traffic-gen-ui
      dockerfile: Dockerfile
    network_mode: host
    # Serves on port 5002 (host network, defined in nginx.conf)

  traffic-gen-responder:
    restart: unless-stopped
    container_name: obmp-traffic-gen-responder
    healthcheck:
      test: ["CMD-SHELL", "bash -c 'echo > /dev/tcp/localhost/5053'"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 30s
    profiles: ["test"]
    mem_limit: 1g
    build:
      context: ./traffic-gen
      dockerfile: Dockerfile
    cap_add:
      - NET_RAW
      - NET_ADMIN
    environment:
      - TRAFFIC_GEN_PORT=5053
      - TRAFFIC_GEN_MODE=responder
      - TRAFFIC_GEN_RESPONDER_MODE=echo
      - TRAFFIC_GEN_INTERFACE=eth0
    networks:
      traffic-test-net:
        ipv4_address: 172.30.0.10
    ports:
      - "5053:5053"

  # GoBGP -- pulls the full real Internet routing table (roadmap E1) from the
  # AS57355 lab route server and BMP-exports it to the OpenBMP collector, where
  # it lands in PostgreSQL ip_rib as a monitored peer. Config + MRT fallback
  # script live in ./gobgp (see gobgp/README.md). Receive-only, local AS 65001.
  gobgp:
    restart: unless-stopped
    container_name: obmp-gobgp
    image: jauderho/gobgp:v4.5.0
    # Host networking: the daemon uses the host's real IPv4 + IPv6 stack, so
    # both the v4 and v6 eBGP sessions to AS57355 source from the host's
    # public addresses (no Docker IPv6/NAT plumbing). BMP still reaches the
    # collector on 10.40.40.202:5000 (its published port).
    network_mode: host
    depends_on:
      - collector
    # gobgpd reads /config/gobgpd.conf; the same mount carries mrt-refresh.sh
    # and the cached MRT dumps it downloads.
    volumes:
      - ./gobgp:/config
    command: ["gobgpd", "-f", "/config/gobgpd.conf", "-t", "toml"]

  # GoBGP -- modular EVPN test-route injector (roadmap E5). Profile-gated, so
  # it is NOT part of the normal stack. Originates synthetic BGP EVPN routes
  # and BMP-exports them so the EVPN pipeline can be exercised. Start only for
  # testing:  docker compose --profile evpn-test up -d gobgp-evpn
  # then:     bash gobgp-evpn/inject-evpn.sh
  gobgp-evpn:
    restart: unless-stopped
    container_name: obmp-gobgp-evpn
    profiles: ["evpn-test"]
    image: jauderho/gobgp:v4.5.0
    depends_on:
      - collector
    volumes:
      - ./gobgp-evpn:/config
    command: ["gobgpd", "-f", "/config/gobgpd.conf", "-t", "toml"]

  whois:
    restart: unless-stopped
    container_name: obmp-whois
    healthcheck:
      test: ["CMD-SHELL", "bash -c 'echo > /dev/tcp/localhost/43'"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 30s
    image: openbmp/whois:2.2.0
    mem_limit: 1g
    sysctls:
      - net.ipv4.tcp_keepalive_intvl=30
      - net.ipv4.tcp_keepalive_probes=5
      - net.ipv4.tcp_keepalive_time=180
    ports:
      - "4300:43"
#    volumes:
#      - ${OBMP_DATA_ROOT}/config:/config
    environment:
      - POSTGRES_PASSWORD=openbmp
      - POSTGRES_USER=openbmp
      - POSTGRES_DB=openbmp
      - POSTGRES_HOST=obmp-psql
      - POSTGRES_PORT=5432

  authelia:
    restart: unless-stopped
    container_name: obmp-authelia
    profiles: ["auth"]
    mem_limit: 256m
    image: authelia/authelia:4.38
    ports:
      - "9091:9091"
    volumes:
      - ${OBMP_DATA_ROOT}/authelia:/config
    environment:
      - TZ=UTC

  portal:
    restart: unless-stopped
    container_name: obmp-portal
    healthcheck:
      test: ["CMD-SHELL", "wget -q --spider http://localhost:80/ || exit 1"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 20s
    profiles: ["auth"]
    mem_limit: 128m
    image: nginx:alpine
    ports:
      - "8080:80"
    volumes:
      - ./portal:/usr/share/nginx/html:ro

networks:
  traffic-test-net:
    driver: bridge
    ipam:
      config:
        - subnet: 172.30.0.0/24