Improve how the test framework determines if the consensus node has shut itself down

Right now, the way an Otter test determines if a consensus node has shut itself down is to send a ping to the `DockerManager`. This ping is initiated from the test. It would be better if the `ContainerNode` detects this automatically, or if the `DockerManager` monitors the process and tells the `ContainerNode` that the node has died so it can update the platform status and lifecycle accordingly.

Some discussion is probably necessary to determine the best approach.

Start by reading the Otter test flow involving ContainerNode and DockerManager, especially how the test-initiated ping determines shutdown. Clarify the monitoring and lifecycle requirements before choosing an approach; done means a consensus node's self-shutdown is detected automatically and the ContainerNode platform status and lifecycle are updated without the test initiating the ping.

Written by the indexing model from the issue text.