Safe shutdown

Please kindly implement way to handle 'safe shutdown' of knotx process.

By 'safe shutdown' I understand waiting until all requests are served or timed out and only after that shutting down process.

The issue does not name any files, tests, or entry points. Start by tracing the process lifecycle and request handling, then identify where shutdown is initiated and how in-flight requests are tracked. Done means the process waits for requests to finish or time out before exiting, with coverage for both completion paths.

Written by the indexing model from the issue text.