Restart container if its application hangs

Question

I have a critical container application which task consist of polling sensors and taking consequent action on some external hardware.

If the application of the container crashes it's fine. restart: always will take care of this.

However if the application hangs, possibly blocked on IO or another unexpected event, this is where things can go wrong as the external hardware will get stuck in it's state possibly indefinitely.

I'd like my container application to periodically communicate with something and if it stops doing so, the container gets restarted.

Does this kind of mechanism exists?

I'm running the container with Docker.

Answer 1

by using docker stack you can achieve a health check. you can add the below lines in docker-compose.yaml.

version: '3.1'

services:
  web:
    image: docker-flask
    ports:
      - '5000:5000'
    healthcheck:
      test: curl --fail -s http://localhost:5000/ || exit 1
      interval: 1m30s
      timeout: 10s
      retries: 3

Reference: https://howchoo.com/devops/how-to-add-a-health-check-to-your-docker-container

Answer 2

Create health endpoint inside application and call it from outside and if no response restart docker