Add a watcher with a maximum timeout to kill containers

As the main purpose of this tool is to spawn new instances that are usually used in the CI, it can be helpful to have a watcher that verifies the running containers and kill any container with an up time that is considered too long for the CI.

1h seems to be a good initial value, this must be set by environment variable or CLI using `clap`.

depends on #1.
