A five-minute check can miss a two-minute outage
If a service passes at 12:00, fails at 12:01 and recovers at 12:03, checks scheduled at 12:00 and 12:05 can both pass. A one-minute schedule provides more observation points. The exact outcome depends on timing, request duration and the checker’s schedule.
Detection and notification are different
An outage may start just after a successful check. The next scheduled request can take time to complete or reach its timeout. Confirmation rules, processing and the notification channel add time. Do not treat “one-minute checks” as a promise that every alert arrives within 60 seconds.
How Statance schedules checks
Free supports a minimum five-minute interval from one region. Pro supports a minimum one-minute interval from up to three regions. Business supports a minimum one-minute interval from up to six regions. Due checks are scheduled by a minute-based job. A monitor can use a longer interval.
For non-heartbeat checks, Statance evaluates the returned regional results and marks the check as failing when more than half of those results fail. A missing probe result is not the same as a successful check. With a single region, that region’s failure is enough to fail the check. The current scheduler transitions monitor state after that result evaluation; it does not expose a separate configurable consecutive-failure or consecutive-success threshold for these checks.
Choose an interval by the consequence of failure
- For a side project, five-minute checks can be a reasonable starting point if delayed detection is acceptable.
- For a production API or checkout, use one-minute checks and test your alert channel.
- For requirements below one minute, evaluate a service that explicitly supports that frequency. Statance does not currently provide sub-minute checks.
Rehearse a controlled failure and recovery. Record observed check, incident and delivery timestamps rather than relying only on a plan label.
See current plan limits or combine uptime with server metrics.