Fail topology wait strategy if container is not running

[npepinpe]

Describe the bug

It can happen (e.g. OOM killed) that the container stops while we're checking the topology. In such cases, right now, the container tries until it times out. We should instead check on each try if the container is still running, and if not simply exit.

To Reproduce

Start a container with too little memory with the topology check.

Expected behavior

Exit early of the topology check if the container stopped running.