Add process monitoring and graceful shutdown handlers

[tobie]

Summary

This PR adds comprehensive process monitoring and graceful shutdown capabilities to the application, including memory usage tracking, process event handlers, and improved logging for debugging production issues.

Key Changes

• Memory monitoring: Added memory() function to capture process memory usage (RSS, heap, external, array buffers) in MB for inclusion in log lines
• Process event handlers: Implemented installProcessHandlers() to catch and log uncaught exceptions, unhandled promise rejections, process warnings, and shutdown signals (SIGTERM, SIGINT, SIGHUP)
• Graceful shutdown: Added signal handlers that trigger server shutdown with a 5-second timeout to prevent hanging processes
• Request-level logging: Added debug-level logging for incoming requests with memory snapshots, excluding health checks and loopback traffic to reduce noise
• Periodic memory snapshots: Added 60-second interval memory usage logging to help identify memory leaks in production
• Server error handling: Added error handler for server startup failures
• Enhanced startup logging: Improved startup log to include memory usage and startup duration

Implementation Details

• Memory values are rounded to nearest MB for readability
• Loopback requests (127.0.0.1, ::1, ::ffff:127.0.0.1) and /health endpoint are excluded from request logging to reduce log noise
• Shutdown timeout of 5000ms prevents indefinite hangs if connections don't close gracefully
• All process handlers log with memory snapshots for correlation with resource usage
• Memory snapshots use unref() on intervals to prevent them from keeping the process alive

https://claude.ai/code/session_01CnAkjjVdTENcYvZSiGWWVA