Logging events when fdbserver gets a kill signal from the operating system

When an fdbserver process is killed through the CLI, we log a trace event. This can be helpful when analyzing the timing of process restarts. If a process is killed through an OS kill signal, we don't have any logging until it starts back up. This can be a hindrance when debugging rolling bounces on Kubernetes, where the container is getting killed through a standard kill signal. I think we should add an event when we get a soft kill signal to help with debugging these kind of scenarios.