Fix/runtime lifecycle cleanup

[shayam137-glitch]

What does this PR do?

This PR tightens the runtime lifecycle and queue behavior around startup failures and worker restarts.

• Rolls back database, queue, watcher, and router resources when plugin initialization fails so startup errors do not leak timers, workers, or open DB handles.
• Preserves the original plugin error instead of masking it with a later failure.
• Makes repeated queue start() calls idempotent for the same worker and rejects a different active worker callback to avoid silent callback replacement.
• Defines explicit behavior for unknown queue job types by failing loudly instead of silently dropping them.
• Adds focused regression tests covering plugin rollback, restart semantics, and queue safety.

How to test?

1. Run the full project test suite:
   • bun test
2. Confirm the lifecycle and queue coverage passes, especially the focused checks for:
   • failed plugin init rollback
   • queue worker replacement rejection
   • queue stop/restart processing behavior
   • unknown job type dispatch failure

Checklist

• My code follows the code style of this project.
• I have added tests for my changes.
• I have updated the documentation accordingly.
• I have run bun run test locally.
• I have run bun run lint locally.

Issue Reference

Closes #468
Closes #470
Closes #471
Closes #473

[drips-wave]

@shayam137-glitch 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