Documentation

Triple Pat watches the watchers. A quiet alerting system cannot tell you whether it is quiet because everything is fine or because it is broken, so you have it fire a synthetic alert on a schedule, with a Triple Pat timer as the alert’s notification target. Every delivered alert is a check-in, and if the check-ins stop, the app on your phone raises an alarm. Everything you need to set that up is here, organized by what you are trying to do: learn the basics, get a specific task done, look up a detail, or understand how the service works under the hood.

Tutorials

Learning-oriented lessons that take you from nothing to a working check-in

• Your first check-in — Send a check-in to a real Triple Pat timer and watch the service record it
• Checking in from anywhere — Read a timer, check in over HTTP through different servers and TLDs, and check in by email

How-to guides

Directions for getting specific tasks done

• Check in from the Watchdog alert — Route the always-firing Watchdog alert your Prometheus stack already ships to a Triple Pat timer, so an alarm fires if your alerting pipeline breaks
• Check in from a cron job — Add a Triple Pat check-in to a script that cron already runs on a schedule

Reference

Technical facts: UUIDs, URLs, and the check-in API

• Check-in API — Timer UUIDs, the check-in URL format, and the public test UUID
• API documentation (OpenAPI) — Swagger-generated reference for every API endpoint

Explanation

How the service is built, and why it is built that way

• One mechanism, many names — Watchdog timers, heartbeats, dead man's switches, snooze buttons, fire drills, and check-in timers are mostly the same idea. How the names map, and why we say check in.
• Design of the checkin service — The design of the Triple Pat checkin and monitoring service
• What makes Triple Pat reliable? — What reliable means for a check-in, how we are set up for success, and our track record