Independent project
Dead Man Ping
A dead man switch for cron jobs and backups: one curl line at the end of a script you already run.
Problem
A backup or cron job can skip a night, write an empty file, or hang, and nothing notices until someone looks.
Approach
Stay out of job execution. Watch pings, intervals, grace periods, optional timeouts, and payload rules configured in the UI.
Solution
Create a monitor in the dashboard, ping it from the existing script, optionally send payload fields, and get alerts on state changes instead of a firehose.
Architecture / technical highlights
Dashboard plus HTTP ping endpoints. Validation rules live in the monitor, not in an SDK. Internal infrastructure is not documented here.
Key features
- Ping from an existing cron or backup script
- Optional payload validation (for example file size)
- Grace period and late / missed job detection
- Optional max execution time
- Alerts on state transitions
Result
A live app on Vercel with sign-in, docs, and a public product site. No customer or volume metrics published here.
Current state
Live at dead-man-ping.vercel.app.
Next steps
Keep the monitoring product shipping. A dedicated domain can replace the Vercel URL when it exists.