feat: add resilient background job retry and monitoring system

[zxzok]

Summary

Implements a resilient background job execution system with automatic retry, exponential backoff, and a monitoring dashboard.

Changes

Backend

• New JobExecution model — tracks job lifecycle (pending → running → completed/failed/retrying)
• New services/jobs.py — job enqueue, execute with retry (exponential backoff: 30s → 120s → 480s), process pending jobs, stats aggregation
• New routes/jobs.py — 4 endpoints: GET /jobs/stats, GET /jobs/recent, POST /jobs/retry/<id>, POST /jobs/process
• Updated services/reminders.py — wraps send_email/send_whatsapp with job system, backward compatible
• 17 tests covering retry logic, backoff timing, max attempts, stats, and all endpoints

Frontend

• New api/jobs.ts — API client module
• New JobMonitor.tsx — monitoring dashboard with stats cards, status filters, recent jobs table, retry button
• Route added at /jobs

How It Works

1. enqueue_job("reminder_email", payload) creates a pending job record
2. execute_job(job_id) runs it via registered handlers
3. On failure: status → "retrying", next_retry_at set with exponential backoff
4. After max_attempts (default 3): status → "failed"
5. process_pending_jobs() picks up retryable jobs (call from external cron or /jobs/process)

Test plan

• Job creation and execution flow
• Retry logic with exponential backoff
• Max attempts exceeded → failed
• process_pending_jobs finds eligible jobs
• Stats endpoint returns correct counts
• Manual retry of failed jobs
• Reminder integration with mocked send functions