Glossary
Check interval
A check interval is how often a monitor checks your service, such as every minute or every five minutes. It caps how fast an outage can be noticed.
Also called Check frequency, Polling interval.
The check interval, also called check frequency or polling interval, is the time between two checks of the same monitor. At a one-minute interval a service is probed sixty times an hour; at five minutes, twelve. It sets a hard floor on detection time: a service checked every five minutes cannot be noticed as down in less than five minutes, and on average will take about half an interval longer than the outage itself to register. Everything else in the alerting pipeline is fast by comparison, so this is the number that decides how quickly you can possibly know.
It also sets the resolution of your uptime figure. Short outages between checks are invisible; outages that straddle a check are rounded to interval boundaries. At a five-minute interval, a 99.99% target is not really measurable — a single failed check is already several times the monthly budget.
Shorter is not free. Every interval is load on the service being checked, and an aggressive interval against a rate-limited or expensive endpoint can cause the failure it is watching for. The usual answer is to check the customer-facing path frequently and the expensive internal checks rarely, rather than setting one interval for everything.
One interaction is worth planning for: if a monitor requires several consecutive failures before it alerts, the interval multiplies. Two failed checks at one minute is a two-minute delay; two failed checks at five minutes is ten. Choose the pair together rather than separately.
Questions
What check interval should I use?
One minute for customer-facing websites and APIs, where every undetected minute is a minute of customers finding the outage first. Five minutes or longer for internal, expensive or rate-limited endpoints, where a tight interval adds load for little benefit.
Is the check interval the same as the timeout?
No. The interval is how often a check runs; the timeout is how long a single check waits for an answer before it counts as failed. In Uptimely the request timeout defaults to 30 seconds and can be set from 1 to 60 seconds on website and API monitors.
How does the check interval affect alerting?
Alerts usually wait for a failure threshold, which is counted in checks. Two consecutive failed checks at a one-minute interval means roughly two minutes before an incident is declared; at a five-minute interval it is roughly ten.
What is the shortest check interval in Uptimely?
One minute, on every plan including the free one.
In Uptimely
Uptimely allows a one-minute interval on every plan, including the free one, and monitors default to declaring an incident after two consecutive failed checks. Both are per-monitor settings.
Uptimely checks every minute on every plan, including the free one, and each monitor keeps its own interval.
Start monitoring free →Read next
Related terms
Part of the Uptimely monitoring glossary — plain definitions of the vocabulary uptime monitoring, incident response and status pages are discussed in. Something here read wrong to you? Email support@getuptimely.com.