Monitor Docker Swarm with TICK InfluxData Stack

docker-compose-tick.yml

version: '3'
services:
  # FRONT
  chronograf:
    # Full tag list: https://hub.docker.com/r/library/chronograf/tags/
    image: chronograf
    deploy:
      replicas: 1
      placement:
        constraints:
          - node.role == manager
      restart_policy:
        condition: on-failure
    volumes:
      # Mount for chronograf database
      - chronograf-data:/var/lib/chronograf
    ports:
      # The WebUI for Chronograf is served on port 8888
      - "8888:8888"
    networks:
      - influx
    depends_on:
      - kapacitor
      - influxdb
  # MIDDLE
  kapacitor:
    # Full tag list: https://hub.docker.com/r/library/kapacitor/tags/
    image: kapacitor
    deploy:
      replicas: 1
      placement:
        constraints:
          - node.role == manager
      restart_policy:
        condition: on-failure
    volumes:
      # Mount for kapacitor data directory
      - kapacitor-data:/var/lib/kapacitor
      # Mount for kapacitor configuration
      - /etc/kapacitor/config:/etc/kapacitor
    ports:
      # The API for Kapacitor is served on port 9092
      - "9092:9092"
    networks:
      - influx
    # Kapacitor requires network access to Influxdb
    depends_on:
      - influxdb
  # BACK
  telegraf:
    # Full tag list: https://hub.docker.com/r/library/telegraf/tags/
    image: telegraf
    environment:
      HOSTNAME: "telegraf"
    deploy:
      mode: global
      restart_policy:
        condition: on-failure
    volumes:
      # Mount for telegraf configuration
      - /etc/telegraf:/etc/telegraf
      # Mount for Docker API access
      - /var/run/docker.sock:/var/run/docker.sock
    networks:
      - influx
    depends_on:
      - influxdb
  # DATABASE
  influxdb:
    # Full tag list: https://hub.docker.com/r/library/influxdb/tags/
    image: influxdb
    deploy:
      replicas: 1
      placement:
        constraints:
          - node.role == manager
      restart_policy:
        condition: on-failure
    volumes:
      # Mount for influxdb data directory
      - influxdb-data:/var/lib/influxdb
      # Mount for influxdb configuration
      - /etc/influxdb/config:/etc/influxdb
    ports:
      # The API for InfluxDB is served on port 8086
      - "8086:8086"
    networks:
      - influx
networks:
  influx:

volumes:
  chronograf-data:
  kapacitor-data:
  influxdb-data:

(from https://gist.github.com/cdelaitre/85949d8b697359a319e30a678e23d8bd#gistcomment-2131758)

Requirements

• a docker swarm with n managers, p workers

1/ Prepare configuration files

• on managers :
  • create /etc/kapacitor/config/kapacitor.conf from https://github.com/influxdata/sandbox/blob/master/kapacitor/config/kapacitor.conf
  • create /etc/influxdb/config/influxdb.conf from
    https://github.com/influxdata/sandbox/blob/master/influxdb/config/influxdb.conf
• on managers & workers :
  • create /etc/telegraf/telegraf.conf from
    https://github.com/influxdata/sandbox/blob/master/telegraf/telegraf.conf
  • In telegraf.conf replace $HOSTNAME with the name of the node (otherwise it will be "telegraf" for all nodes)

2/ Copy and Run the stack

• on a manager :
  • copy the above file docker-compose-tick.yml
  • run command :

docker stack deploy --compose-file docker-compose-tick.yml tick

3/ Set up

• Go to Chronograf URL : http://<LAN_IP_OF_A_NODE>:8888
• Homepage > Set Connection String to http://influxdb:8086
• Menu ALERTING > Configure Kapacitor > set URL to http://kapacitor:9092

Caution

• influxdb and kapacitor components are not clustered

Since remote cfg is available for telegraf, how to avoid hardcoding of node names?