Ansible zero downtime for docker_container

site.yml

  vars:
    version: "{{ ansible_date_time.epoch }}"
    backend_container_name: "backend-{{ version }}"
    
  tasks:
   - name: Run backend container
      docker_container:
        <<: *base-app
        name: "{{ backend_container_name }}"
        command: bundle exec puma -C config/puma.rb -t 1:3
        restart_policy: always
        healthcheck:
          test: ["CMD", "sh", "-c", "wget -S localhost:3000/shop 2>&1 | grep -q '200 OK'"]
          interval: 15s
          timeout: 2s
          retries: 3
          start_period: 15s
        labels:
          traefik.enable: "true"
          traefik.port: "3000"
          traefik.A.frontend.rule: "Host:xxxx;PathPrefix:/shop"

    - name: Wait backend healthy
      shell: "docker ps | grep {{ backend_container_name }}"
      register: ps_cmd
      retries: 10
      delay: 15
      until: ps_cmd.stdout_lines | reject('match', ".*\(healthy\).*") | list | count == 0

    - name: Stop other containers
      shell: "docker ps | grep backend | grep -v {{ backend_container_name }} | awk '{ print $1 }' | xargs docker stop | xargs docker rm"