Cron monitoring guide

Know when a scheduled job doesn't finish.

Your job checks in when it runs. CronRadar watches the schedule and tells you when a run fails, takes too long, or never arrives. Start with one job and verify its next scheduled run.

Start free Choose an integration

Missed-run alerts
Start and finish history
Group jobs by application

Runnable setup

Report a start, a finish, or a failure

Create an application and store its key as CRONRADAR_API_KEY in the job's environment. Save this wrapper as a script and replace the backup command. This example expects a daily run at 02:00 UTC; match the real job's schedule and timezone in your monitor settings.

BashCopy, add your key, run

#!/usr/bin/env bash
# The backup keeps its own exit status if a check-in fails.
base="https://cron.life/api/v1/ping/nightly-backup"
run_id="$(date -u +%Y%m%dT%H%M%S)-$$-$RANDOM"

report() {
  if [ -z "$CRONRADAR_API_KEY" ]; then
    printf '%s\n' 'CronRadar key missing; backup will still run.' >&2
    return 0
  fi
  curl --silent --show-error --fail --max-time 5 \
    --request POST --user "$CRONRADAR_API_KEY:" \
    --get "$base/$1" \
    --data-urlencode 'schedule=0 2 * * *' \
    --data-urlencode 'gracePeriod=300' \
    --data-urlencode "run_id=$run_id" \
    --output /dev/null || \
    printf '%s\n' 'CronRadar check-in failed; inspect monitoring setup.' >&2
}

report start
/opt/app/backup.sh
job_status=$?
if [ "$job_status" -eq 0 ]; then
  report complete
else
  report fail
fi
exit "$job_status"

Each report can wait up to five seconds. A failed report does not stop the backup or replace its exit status. The first accepted report registers the job; create the monitor in advance if you need missed-run coverage before that first execution.

Follow the setup guide to connect it, or give your assistant the agent quickstart.

What you can track

The schedule, the result, and what needs attention.

A missing run

If the machine, scheduler, worker, or job disappears, the expected CronRadar signal disappears with it.

A late start

CronRadar evaluates the next expected run against a bounded grace period instead of treating every job like a fixed uptime probe.

Starts and finishes

Send start, complete, fail, and a bounded failure message. Keep detailed logs in the system built for them.

Jobs by application

Group recurring work by application while framework extensions keep declared schedules synchronized.

Alerts where you work

Route incidents to email, Slack, webhook, Discord, or PagerDuty on every plan.

Keep your jobs running

Official SDKs cap HTTP requests at five seconds and catch monitoring failures so the job keeps its original behavior.

Real interface

See protected work by application, not by ping URL.

CronRadar workspace showing applications with healthy, warning, and critical automation counts — The application view groups jobs and shows healthy, late, and failed runs. Example data is shown.

Verify your setup

Check a real run before relying on alerts.

Connect one job

Choose a job whose output you can safely inspect. Use an available framework integration or the shell wrapper above.

Check the run and alert

Confirm a real completion and its output. Send a test notification, then use a dedicated test job to verify failure or missed-run behavior.

Watch the next schedule

Confirm the next scheduled execution arrives. Keep existing monitoring active until the schedule and alert destination are verified.

Questions

Focused monitoring has deliberate limits.

Is this uptime monitoring?

No. CronRadar receives signals from scheduled work; it does not probe websites or publish status pages.

Does it collect logs or traces?

No. Store diagnostic detail in your logging platform. CronRadar keeps execution events and an optional bounded failure message.

Can it detect a job that ran but processed zero records?

Yes. Define a bounded expectation such as records_processed >= 1, then report that numeric result with the completion event. Missing or failing checks create an explained outcome incident.

Where should I start?

Choose your framework in the setup guide and check its current package availability. Direct HTTP reporting works when an SDK is not available.

See the complete workflow on the product overview.

Give one important automation an independent signal.

Free protects ten active automations with unlimited collaborators and every supported alert channel. Development and staging jobs count toward the same limit.

Start free See pricing