docker-compose의 로그를 혼잡하게 CloudWatch에 던지기

경위

환경

docker-compose.yml

logging:
  driver: "awslogs"
  options:
    awslogs-region: "ap-northeast-1"
    awslogs-group: "${RAILS_ENV}"
    awslogs-stream: "nginx"

Dockerfile

RUN ln -sf  /dev/stdout /var/log/HOGOHOGE/app-access.log
RUN ln -sf  /dev/stderr /var/log/HOGEHOGE/app-error.log

cloud watch

빈 greoup를 만들지 않으면 다음 오류가 발생합니다.

ERROR: for redis  Cannot start service redis: Failed to initialize logging driver: ResourceNotFoundException: The specified log group does not exist.
    status code: 400, request id: xxxxxx-xxxxxxxxx-xxxxxxxx-xxxxx

ERROR: for db  Cannot start service db: Failed to initialize logging driver: ResourceNotFoundException: The specified log group does not exist.
    status code: 400, request id: xxxxxx-xxxxxxxxx-xxxxxxxx-xxxxx

움직여 보자

$ docker-compose up

로그 집계 가능

docker-compose.yml 전체 (db와 redis에는 의미가 없습니다)

version: '2'
services:
  web:
    image: nginx
    depends_on:
      - db
      - redis
    ports:
      - "80:80"
    logging:
      driver: "awslogs"
      options:
        awslogs-region: "ap-northeast-1"
        awslogs-group: "develop"
        awslogs-stream: "nginx"
  redis:
    image: redis
    logging:
      driver: "awslogs"
      options:
        awslogs-region: "ap-northeast-1"
        awslogs-group: "develop"
        awslogs-stream: "redis"
  db:
    image: postgres
    logging:
      driver: "awslogs"
      options:
        awslogs-region: "ap-northeast-1"
        awslogs-group: "develop"
        awslogs-stream: "posgre"

요약