Check-in reliability
Seiri’s reliability model is push-based: your job must check in. Seiri alerts when the check-in is late or reports failure.
This is not outbound webhook delivery / retries to your servers.
What Seiri guarantees
- Accept check-ins on
ping.seiri.app(HTTP) or{ref}@ping.seiri.app(email). - Evaluate schedules for Periodic and Cron monitors against the grace period.
- Alert on miss / fail via your configured channels.
- Retain history for dashboard, webhook stats, SLA, and status-page automation.
What you must do
- Ping after successful work (
/success). - Optionally ping
/failon errors. - Keep ping timeouts short and swallow ping errors so monitoring never takes down the job.
- Set grace period above your job’s p99 runtime to avoid flapping.
Missed vs failed
| Signal | Meaning |
|---|---|
| Missed | No check-in arrived before schedule + grace |
| Failed | Job called /fail (or email subject [FAIL] / [ERROR]) |
| Healthy | Check-in arrived on time as success |
See Ping API and Integrate a job.