Stopping docker container exceeds grace period

[wvffle]

When I stop my docker container discord, steam and slack (basically all of mx-puppet-bridge based) bridges are exceeding the grace period defined in the docker-compose.yml every single time.

Setting the stop_grace_period to a low value does not satisfy me as those bridges are clearly not shutting down correctly.
Logs do not contain any errors so probably the node server does not receive the SIGTERM

I have following docker-compose.yml:

version: '2.4'
services:
  discord:
    build: https://github.com/matrix-discord/mx-puppet-discord.git
    restart: always
    stop_grace_period: 2m
    volumes:
      - /opt/mx-puppet-discord:/data
    ports:
      - 8434:8434

[JuniorJPDJ]

App doesn't react on SIGTERM, it should shutdown itself if this signal is coming ;/

[gyroplast]

Consider executing the container process in an init system by adding init: true to the compose options, this will properly forward signals to the application and terminate the container gracefully within seconds as expected.

See https://docs.docker.com/compose/compose-file/compose-file-v2/#init for reference.