Automations
Uptime and health check automation
Every 30 minutes: your endpoints checked, incidents logged, silence when all is well.
Last reviewed July 17, 2026
A cron automation recipe that probes your public endpoints on a schedule with code execution, logs results, and reports only on failures.
Schedule: Every 30 minutes
*/30 * * * *Automation instructions
Using code execution, request each URL in /Ops/uptime-targets.md (one per line: URL, expected status, must-contain string if any, timeout seconds). For each: record status code, response time, and content check result. Append one summary line per run to /Ops/uptime-log.md (timestamp, pass count, fail count). ONLY when a target fails: add a dated incident entry to /Ops/uptime-incidents.md with the target, the observed failure, the response body's first lines, and whether the previous run also failed (repeat failures get escalating markers). No prose reports on healthy runs; the log line is enough.
Set it up
- 1
List your targets
Create /Ops/uptime-targets.md: your site, API health endpoint, docs, anything whose silence would embarrass you.
- 2
Copy the instructions
The incidents file is designed to be greppable history; the log file answers 'since when'.
- 3
Create the automation with a small budget
Every 30 minutes at a few HTTP checks per run stays cheap; the budget cap keeps it provably so.