Skip to content

Cron job monitoring · automatic discovery

Keep your cron jobs on the radar.

Get alerted when a job misses its schedule, fails, or gets stuck. See what was expected, what happened, and when it last worked.

No credit card · 10 monitors free

Sample data. A radar sweep checks in on scheduled jobs. nightly-backup was due at 02:00 UTC and sent no ping within its 5 min grace period, so an example email alert went out. invoice-run reported a failed run, so an example Slack alert went out. weekly-report started at 04:00 and did not finish within 1 h, so an example webhook marked it stuck. On the next sweep all three were healthy again.

Works with your scheduled jobs

all →

Discovery for Hangfire, Quartz.NET, Celery and Laravel. HTTP pings for any other scheduler.

Set up with your AI assistant

Choose a workspace, copy your setup prompt, and paste it into your coding assistant. It will identify your scheduler and guide the integration.

Connect your assistant Read the setup steps

How discovery works

Install the package. The jobs you already have show up on their own.

The extension reads your framework's own schedule, so names and schedules are never declared twice. An API key is the only required setting, and every job is monitored unless you opt it out.

C# · setup
dotnet add package CronRadar.Hangfire

// Program.cs, after AddHangfire and AddHangfireServer
builder.Services.AddCronRadarHangfire();
Discovered jobsexample
  • nightly-backup0 2 * * *healthy
  • invoice-run*/15 * * * *healthy
  • weekly-report0 6 * * 1 · stopped runningcritical
Found from recurring jobs in the Hangfire registry, with names and schedules taken from your code. No per-job setup.

What you hear about

The jobs that stop running, vanish or never ran.

Each one has its own mark and its own message, so you know which kind of silence you are looking at.

Stopped running

critical

The expected run passes, the grace period ends, and no report arrives. The job turns critical and the alert goes out.

example · weekly-report · no run since Mon 06:00

Vanished

paused

A job removed from the framework’s registry sends one alert and is paused, so it doesn’t alert forever. A deploy that drops every job sends one alert, not one per job.

example · sync-crm · left the registry · alert sent

Never ran

critical

A job that was found but never reported by its first expected run is flagged, so a schedule that never fires doesn’t look healthy.

example · cleanup-temp · no first report by 03:30

Beyond did-it-run

Did it run, and did it run correctly?

Everything here is about scheduled work. No uptime checks, status pages or log storage.

Stuck runs

A run that reports its start but never its finish turns critical, so a hung job doesn’t sit there looking busy.

start → no completion

Outcome expectations

A run can report up to five numbers or true/false values. Expect rows_exported ≥ 1, and a run that finishes but breaks it is reported.

outcomefailed

One alert, then a daily reminder

You get one alert when a job turns critical, then a reminder each day until someone acknowledges it or the job recovers.

InitialFailure → ContinuedFailure → Recovered

Five alert channels

email, Slack, webhook, Discord, and PagerDuty, on every plan including Free. Hand off to PagerDuty when you need on-call.

no per-channel charge

Agent and MCP setup

Give your coding agent a private setup prompt with temporary workspace access, or connect the hosted MCP server.

See agent setup

Monitoring never breaks the job

SDK requests time out after 5 seconds, monitoring errors are caught, and your job’s own exceptions are re-thrown unchanged.

timeout: 5s

Agent setup

Hand the integration to your coding agent.

Most teams now ask an agent to wire up monitoring. CronRadar gives it a path it can follow on its own, with permissions you choose first.

  • A private setup prompt. Pick a workspace and its access level; the prompt carries a temporary token you can revoke.
  • A hosted MCP server. api.cronradar.com/api/agent/mcp lets an agent create jobs, read check-ins and pause or resume monitoring with a bearer token.
  • Docs agents can read. llms.txt and a Markdown version of every docs page, plus the agent quickstart.

Pricing

Free for your first 10 monitors.

Then $1 per monitor a month, whatever the schedule. Unlimited teammates, every alert channel on every plan, and an optional monthly budget.

FAQ

Questions, answered.

Put your scheduled jobs on the radar.

10 monitors free, no credit card. Install one package and see what it finds.