fix(#149): drain HTTP, Socket.IO, and DB work on graceful shutdown

[spiffamani]

Summary

• Centralize SIGTERM/SIGINT handling in GracefulShutdownService so the server stops accepting new HTTP/Socket.IO work, drains in-flight requests, disconnects Socket.IO clients, waits for tracked MongoDB sessions, then closes the DB pool before exit.
• Remove the competing SIGINT handler in database.ts that could exit before sockets/HTTP finished draining.
• Add request-tracker middleware (503 during drain), tracked DB sessions for transactional work, and unit tests covering the shutdown sequence.

Closes #149

Test plan

• pnpm exec jest tests/gracefulShutdown.test.ts --runInBand (5 tests passed)
• Manually start the server, open a Socket.IO client, send SIGTERM, confirm logs show HTTP close → socket drain → DB disconnect → clean exit
• Confirm a mid-request does not get cut off abruptly when SIGTERM is sent
• Confirm new HTTP requests receive 503 after shutdown begins

[drips-wave]

@spiffamani Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits