feat: resilient background job manager with retry & monitoring

[robotica4us-collab]

Implements a production-ready background job system that wraps APScheduler with automatic retry, dead-letter tracking, and full observability.

Features:

• Configurable exponential backoff retry (per-job RetryPolicy)
• Dead-letter queue for permanently failed jobs with admin reset
• Per-job execution history with timing and error details
• Prometheus metrics (executions, retries, dead-letters, duration)
• Redis-backed state persistence for crash recovery
• Admin API endpoints (/jobs/status, /jobs/dead-letters, /jobs//reset)
• Unauthenticated health check (/jobs/health) for monitoring systems
• Automatic Flask app context injection for job functions

The reminder processing job is wired up as the first managed job, replacing the manual /reminders/run endpoint with automatic processing.

Includes 17 tests covering:

• RetryPolicy exponential backoff and delay capping
• JobState history capping
• Job registration and status reporting
• Dead-letter detection and reset
• Flask integration (health endpoint, admin auth)
• Successful execution resets attempts
• Failing jobs retry then dead-letter
• Transient failures recover on retry
• Redis persistence and crash recovery

Closes #130

Summary

• What changed:
• Why:

Validation

• Frontend lint: cd app && npm run lint
• Frontend tests: cd app && npm test -- --runInBand
• Backend tests: ./scripts/test-backend.ps1
• Updated docs if needed

Security and Ownership

• PR opened from a fork (not direct push to main)
• CODEOWNERS review requested

Checklist

• No secrets added
• No unrelated files changed
• Breaking changes documented