Vale construir sua versão de Cronitor?

Build me a cron-job monitor like Cronitor / Healthchecks. Requirements:

- Each monitored job gets a unique ping URL (/ping/:uuid). The job's crontab
  line just appends `&& curl -fsS <url>` · document this.
- Define jobs in SQLite via a basic-auth /admin page: name, expected period
  (e.g. every 60 min) and grace time. Store every ping with timestamp.
- A scheduler loop checks each job: if now > last_ping + period + grace, the job
  is LATE → send one alert webhook (Discord/Slack/Telegram, URL in .env) with
  the job name and how late it is. Alert again on recovery. No alert spam ·
  one per state change.
- /admin dashboard: each job as a row with green/amber/red status dot, last ping
  time ("7 min ago"), and a 24h ping histogram as inline SVG.
- Support /ping/:uuid/fail for explicit failure reports with an optional body
  shown in the alert.
- Python or Node, single process, systemd unit included.