How Cronwize works
Each job calls a unique URL when it runs. When the call does not arrive on time, Cronwize sends an alert.
The model
Cronwize never reaches into your systems. Each monitor has a secret URL, and the job calls it when it runs. If the expected call does not arrive, the silence is the signal.
backup.sh && curl -fsS -m 10 https://ping.cronwize.com/YOUR-KEYSchedules
A monitor expects a ping either on a cron schedule or at a fixed interval.
- Cron in three dialects as written: Unix (5 fields), Spring (6 fields, seconds first) and Quartz (6 or 7 fields, with
?,L,Wand#), each with an IANA time zone. - Interval, such as every 6 hours, from 1 minute to 365 days.
The checker's resolution is one minute, so a schedule that fires more than once a minute is refused with an explanation.
Grace period
The grace period is the time allowed after the expected run before the monitor goes down, from 1 minute to 7 days. Set it longer than the job's normal variation, or the job will alert on a slow night.
Monitor states
| State | Means |
|---|---|
| Down | The expected ping did not arrive within the grace period, a /fail arrived, or a started run did not finish in time. Alerts were sent. |
| Late | The expected time has passed and the grace period is still running. No alert yet. |
| Up | The last expected run pinged on time. |
| New | Created, waiting for the first ping. A new monitor never alerts by deadline. |
| Paused | Records pings, sends no alerts. Resuming is explicit. |
The run strip
Every monitor shows its recent runs as a strip, one tick per expected run, oldest on the left.
Solid green is on time and amber is late inside the grace period. Solid red is a reported failure and a hollow red tick means no ping arrived. Hatching marks a run skipped because Cronwize itself could not receive pings, which the reliability page explains. A blue outline is a run in progress.
Alerts
Alerts go to email, Slack, a signed webhook and mobile push, on every plan. A down alert carries the expected time, the last ping and the first 500 characters of the last failure output, if the job sent any. A recovery alert is sent when a down monitor pings again.
Mobile apps
The Android and iOS apps show monitors by status, monitor detail with the ping log, and pause and resume, and they receive push alerts. They are not in the stores yet. Plans are bought on the web only.