🎉 Feature Request 🎉

What did you expect to see?

Source: As described in this forum thread, I had a missing router, which was not visible in the Dashboard and not routed by Traefik. As Traefik is an ingress controller and can act as reverse proxy, I assumed that this is the scope it would work in. I was quite surprised, that it would also handle a Docker Containers healthcheck result (0/1) and –depending on it– hide a container from the UI and stop routing it. The fact that it routes to a 404 is understandable. But the containers health check gets treated differently from all other errors that can occur: It does not get filed as error in the UI. The container just gets removed from the list of routers.

Luckily I looked at docker-compose ps to wonder if the unhealthy state is connected to this and also stumbled upon a (rare) log entry. The experience I had was the opposite of good. The error was logged only when the health check ran, so this could vary between 10s and 5m, depending on your setup. The message that gets logged notes Filtering unhealthy or starting container. Note: The verbosity could get increased, which is handled in a different feature request, see related.

Feature Request: As long as the container mode is unhealthy or starting, it won't appear in the Dashboard. Please just gray it out and note the container state next to the router. Having this info available would make seeing an unhealthy container and reacting on it, or waiting for a starting container a no brainer instead of the source of frustration and guessing and CLI hacking it currently is.

As a…

developer/ sysadmin/ full-stack-dev-ops-unicorn

I want to…

be informed if a service is unhealthy or starting and soon routable, but not yet ready

so that…

I instantly know if I should just be patient or run and fix an unhealthy container.

Thanks in advance!

Related: #7828