Make flushing to system.crash_log

Why it matters

This feature ensures that when a crash signal is handled by ClickHouse, the logs are flushed synchronously to the system.crash_log table. This guarantees that all crash-related log entries are reliably written and not lost due to asynchronous flushing, improving diagnostics and post-mortem analysis.

How to use it

No explicit user action is required. The synchronous flushing to system.crash_log during signal handling is enabled internally within ClickHouse and works automatically upon receiving crash signals.