[Backend] notificationService cleanup interval lacks a try/catch around its async callback

[Johnpii1]

close #15

Description
Add a guarded helper runScheduledCleanup() that wraps runCleanup() in a void IIFE with try/catch and logs failures via the shared logger.error path.
Replace the immediate startup call and the setInterval callback to invoke runScheduledCleanup() so scheduled failures do not become unhandled rejections and do not stop future runs.
Add unit test coverage that mocks a thrown error from deleteOldNotifications, asserts the error is logged, and verifies a later scheduled run still executes.
Update test scaffolding to mock cacheService and logger and to start/stop the scheduler during test setup/teardown.

Testing
Ran npm test -- --runInBand src/tests/notificationCleanup.test.ts and the test file passed (11 tests passed, 0 failed).
Ran npm run typecheck and the TypeScript check completed with no errors.

[ogazboiz]

clean, merging once ci is green. runScheduledCleanup() now wraps await runCleanup() in try/catch with logger.error and is used for both the startup run and the setInterval callback (was void runCleanup(), which could throw an unhandled rejection and take down the process). a rejection can no longer escape, the interval keeps firing, and errors are logged not swallowed. the new test covering "catch and log without stopping future runs" passes. nice fix.

small thing for next time: this is pushed from your fork's main branch, a descriptive branch (e.g. fix/notification-cleanup-trycatch) is cleaner and plays better with CI association.

if you want to keep contributing, join us on Telegram: https://t.me/+DOylgFv1jyJlNzM0

ps: CI is red right now on the migrate step, but that is a pre-existing main-branch issue (a commonjs migration file under our ESM package), nothing to do with your change. once that is sorted this lands.

[Johnpii1]

Okay

[Johnpii1]

I Will fix the ci