Getting started

A job is a URL, a method and a schedule. The service makes the request on time, records the status code and the timing, and leaves the response body alone. A name labels the row and a description says why the job exists, both kept here and never sent to the target.

Creating one takes a URL and a schedule. Everything else has a working default: GET, a ten second timeout, and no headers. A slow endpoint takes a longer timeout, up to sixty seconds, and a job needing another method has GET, HEAD, POST, PUT, PATCH, DELETE and QUERY to choose from.

A first job

A five minute health check, written as a cron expression:

URL       https://api.example.com/health
Method    GET
Schedule  */5 * * * *   (cron, Europe/Berlin)

The form shows the next five runs as the expression changes, so a wrong field shows up before the job is saved rather than at two in the morning.

What counts as working

A 2xx status is a success. Anything else is a failure, including a redirect, which is recorded and not followed. Three retries follow a failed run, spaced one minute, five minutes and thirty minutes apart. A fourth failure pauses the job and sends one email with a resume link.

Limits

Fifty jobs per account, ten of them on any one host, eight headers per job, a 16 KiB body, and a minimum interval of sixty seconds. The first 24 hours after signup run on smaller numbers: five jobs, two per host, and a five minute minimum. The full limits arrive on their own, with nothing to ask for.