always-up VS docker-autoheal

I developped always-up, a very lightweight service (docker image) that will scan for unhealthy containers and restart them. The main feature is that it handles container dependencies and will restart them in the right order.

autoheal

Have you tried autoheal? It's a Test Driven Design-based "auto"-gpt. It iterates on your code until the tests you've created (or created via chatGPT) are passing. I have had mixed results, but I think having tests serves as a good north star for an agent to work toward.

Hello guys, I often have this issue where a container B uses the network of a container A and container A fails at some point. Also I have many dumb containers that will just fail on their own without being restarted. I tried the nice willfarrell's autoheal but it does not support dependencies. I did not want to migrate my setup to some k3s or something that would be way overkill for what I want.

I use autoheal and have the healthcheck below in my docker compose. I also have uptime-kuma monitoring the web page availability for alerting and Tautulli monitoring remote access availability.

As a workaround for this you could use autoheal to automatically restart the kuma container when it becomes unhealthy.

I useautoheal and i try to have a useful healthcheck on all my containers. It is only lacking any kind of notification if a container is unhealthy and is being restarded.

I've figured a workaround by adding a healthcheck to the docker-compose and running an autoheal container, but I just can't quite figure out why I need it in the first place.

Make sure to implement health hecks, so containers restart automatically when your vpn connection drops. Docker doesn't restart unhealthy containers automatically, but you van use autoheal for that. Also use the health status in the depends_on to make sure the containers start in the correct order (start the *arr containers after the vpn is running for example).