How to Choose a Cron Monitor in 2026: A Buyer's Checklist
Contents
Every cron monitor does the basic job: give it a URL to ping and it alerts when the ping doesn't arrive. The differences that matter show up later: whether it understands your cron schedule or only an interval, whether it can tell a failed or stuck run from a missed one, how much per-job setup it takes, and what the pricing unit does as you add jobs. Pick by your jobs, not by the feature list. This checklist is written by CronRadar, a cron monitor, so weigh it accordingly. Facts about other tools come from their own documentation and pricing pages as checked in November 2026; prices and plans change, so confirm on each vendor's site before you buy. For a tool-by-tool view, see cron monitoring tools compared.
1. Does It Understand Cron Schedules, or Only Intervals?#
An interval check ("expect a ping every 24 hours") is fine for jobs that run at a fixed rate. It breaks for anything with gaps: a job on 0 9 * * 1-5 has a 72-hour gap over the weekend, so a 24-hour interval alerts every Saturday, and a 72-hour interval misses a whole weekday. A monitor that parses the cron expression knows each expected time.
- Cron schedules: Cronitor, Healthchecks.io (cron and systemd
OnCalendar), Better Stack, Sentry Crons, Oh Dear, Hyperping, CronRadar. - Intervals only: Dead Man's Snitch, UptimeRobot heartbeats.
Cron expression syntax covers the fields, and cron schedule shortcuts the @daily family.
2. Which Time Zone Is the Schedule Read In?#
A schedule written for 02:00 in New York means 06:00 or 07:00 UTC depending on the season. If the monitor reads it in UTC and the scheduler in local time, alerts arrive an hour off for half the year. Check whether the tool lets you set a time zone, and at what level: per monitor (Sentry Crons, Oh Dear, Hyperping, WatchCron) or per application (CronRadar, which uses one IANA zone per application, UTC by default). Check what it does with the hour that's skipped or repeated at a daylight saving change. Cron monitor time zones goes into the failure modes.
3. Can It Tell Missed, Failed, and Stuck Apart?#
A single "I'm done" ping detects a run that never happened. It can't detect a run that started and hung, and it reports a failed run only as "late." Look for three signals:
- Start, so duration can be measured and a run that never finishes is caught.
- Fail, so an explicit failure alerts immediately instead of waiting out the grace period.
- Exit code or message, so the alert says what went wrong.
Healthchecks.io supports /start, /fail, and /<exit-code>; Cronitor has run, complete, and fail; Sentry Crons has in_progress, ok, and error with a maximum runtime; Better Stack has /fail and exit codes; Oh Dear has start, finished, and failed with runtime and exit code; Dead Man's Snitch captures exit code, duration, and output through its Field Agent. CronRadar has start, complete, and fail, with stuck-run detection. UptimeRobot heartbeats are fixed-interval pings only. Cron job duration tracking shows why the start signal matters.
4. Can It Check That the Job Did Its Work?#
Exit code 0 doesn't mean the job did anything: an import that found zero rows, or a backup that wrote an empty file, both exit cleanly. Some tools let a run report a value and check it. Cronitor has assertions such as metric.duration < 30 min; CronRadar accepts up to five numeric or boolean outcome values on completion and checks them against expectations you configure, reporting a broken one as outcomefailed. Most heartbeat-only tools don't. Succeeded but did nothing explains the problem.
5. How Much Setup Does Each Job Need?#
This is where tools differ most, and it compounds: the job nobody registered is the one nobody hears about. Three levels:
- Per job, by hand: create a check, copy its URL into the job. Every HTTP-only tool works this way, and it's fine for a handful of crontab lines.
- Per job, in code: an attribute or wrapper on each job, such as Sentry's
[SentryMonitorSlug]for Hangfire or->sentryMonitor()for Laravel. - Discovered from the framework: the tool reads the scheduler's registry and monitors everything.
Discovery support as of November 2026, from each vendor's docs:
| Framework | Discovered automatically by |
|---|---|
| Hangfire | CronRadar (Sentry needs an attribute per job; Cronitor has only a community package) |
| Quartz.NET | CronRadar, Sentry (MonitorAllJobs, SDK 6.13.0 and later) |
| Laravel scheduler | CronRadar, Cronitor (cronitor:sync), Spatie's package with Oh Dear (schedule-monitor:sync) |
| Celery beat | CronRadar, Cronitor, Sentry (monitor_beat_tasks=True) |
| Sidekiq | Cronitor (sidekiq-cronitor), Sentry (sidekiq-cron and sidekiq-scheduler) |
| Kubernetes CronJobs | Cronitor's Kubernetes agent |
| crontab | Cronitor (cronitor sync) |
CronRadar's Node.js, Python, PHP, and .NET base SDKs wrap jobs explicitly; Sidekiq, Kubernetes, and crontab jobs report to CronRadar over HTTP. The framework pages (Quartz.NET, Celery, Laravel) show the setup.
A related question: does the tool create a monitor before the job's first run? A monitor that only appears on the first ping can't tell you the first ping never came. Tools that create monitors from your code ahead of time (Cronitor's sync commands, Spatie's sync, CronRadar's discovery) can.
6. What Happens When a Job Is Removed?#
When a deploy removes a job, a heartbeat check keeps expecting it and eventually alerts "late," or, if nobody created a check, nothing happens. Ask how the tool reconciles: does a sync remove the monitor (Spatie's sync removes Oh Dear monitors no longer in the schedule unless you pass --keep-old), keep it alerting, or pause it with its history (CronRadar.Hangfire pauses monitors whose recurring-job ID disappears)? Detect scheduled jobs that disappear covers why this matters.
7. Where Do Alerts Go, and How Noisy Are They?#
Check that the tool reaches where your team actually responds. Email, Slack, webhooks, and PagerDuty are near-universal. Healthchecks.io lists about 27 integrations including Signal, WhatsApp, and ntfy; Better Stack and Pulsetic include phone and SMS; CronRadar sends to email, Slack, webhooks, Discord, and PagerDuty and hands escalation to PagerDuty. Then check noise controls: grace periods per job, reminders while a job stays down, and recovery messages. Cron job alerting and grace periods cover tuning.
8. Do You Want Only Cron, or a Bundle?#
Most tools bundle cron monitoring with uptime checks, status pages, or on-call: Better Stack, Oh Dear, Hyperping, WatchCron, Pulsetic, UptimeRobot, and Cronitor all include uptime or status pages. That's good value if you need those anyway, and more product to learn if you don't. Dead Man's Snitch, Healthchecks.io, and CronRadar stay focused on jobs. Cron monitoring vs uptime monitoring explains what each kind of check catches.
9. Can Your AI Coding Agent Set It Up?#
In 2026, the person wiring monitoring into a codebase is often a coding agent. Useful signs: an MCP server, an llms.txt file, Markdown versions of the docs, and an OpenAPI description. Cronitor, Sentry, Better Stack, WatchCron, Oh Dear, Hyperping, UptimeRobot, Pulsetic, and CronRadar all offer an MCP server; Healthchecks.io and Dead Man's Snitch don't have an official one. AI agent cron monitoring setup shows what a safe agent setup looks like.
10. What Does It Cost as You Grow?#
Compare the pricing unit, not the entry price, at the job count you expect in a year:
| Tool | Free tier (Nov 2026) | Pricing unit |
|---|---|---|
| Cronitor | 5 monitors | Per monitor plus per user |
| Healthchecks.io | 20 jobs | Plans with a job cap ($20/month for 100) |
| Better Stack | 10 monitors and heartbeats combined | Heartbeat bundles plus responder seats |
| Sentry Crons | 1 monitor | $0.78 per extra monitor on a paid plan |
| Dead Man's Snitch | 1 snitch | Snitch plans ($5/month for 3) |
| Oh Dear | None | Per site (5 cron jobs per site) |
| Hyperping | 20 monitors | Monitor plans |
| CronRadar | 10 monitors | Per monitor per month beyond those; see pricing |
Watch for per-seat charges if many people need access, and check whether development and staging jobs count toward the same limit (with CronRadar, they do). Self-hosting Healthchecks.io or Uptime Kuma costs no license fee but adds a server you have to monitor too; self-hosted cron monitoring weighs that trade.
Quick Picks#
Fair summaries, by situation:
- A few crontab lines, lowest cost: Healthchecks.io's free tier, or self-host it.
- Already on Sentry: Sentry Crons keeps check-ins next to your errors. See the Sentry comparison.
- Uptime, status pages, and cron in one bill: Better Stack, Oh Dear, or Cronitor. See Better Stack and Cronitor.
- Laravel with Oh Dear already in place: Spatie's schedule monitor.
- Simplest possible check-in: Dead Man's Snitch, if intervals suit your jobs. See the comparison.
- Self-hosted framework library: CronWatch. See CronWatch compared.
- Hangfire, Quartz.NET, Celery beat, or Laravel jobs you want covered without per-job setup, in a cron-only tool: CronRadar.
- Coming from Cronhub: see migrating from Cronhub.
Frequently Asked Questions#
What is the difference between a cron monitor and uptime monitoring?#
Uptime monitoring calls your service from outside and alerts when it doesn't answer. A cron monitor waits for your job to report and alerts when the report doesn't arrive on schedule. A website can be up while its nightly billing job hasn't run for a week, which only a cron monitor catches.
Is a free cron monitor good enough?#
Often, for a small number of jobs. Free tiers range from 1 monitor (Sentry Crons, Dead Man's Snitch) to 20 (Healthchecks.io, Hyperping); CronRadar's covers 10. Check the free tier's alert channels and history retention as well as the count, and what the next step costs once you outgrow it.
Do I need start and fail pings, or is one ping enough?#
One ping after success catches missed runs, which is the most common failure. Add a start ping to measure duration and catch runs that hang, and a fail ping so a failure alerts immediately instead of after the grace period. If the job wraps easily, send all three.
Which cron monitors work with Hangfire and Quartz.NET?#
Sentry's .NET SDK monitors Hangfire jobs that carry a [SentryMonitorSlug] attribute and, from version 6.13.0, all Quartz.NET jobs with MonitorAllJobs. CronRadar discovers every Hangfire recurring job and every Quartz.NET triggered job with one call. Other services work through a job filter or listener you write and a check per job. Hangfire monitoring tools and Quartz.NET monitoring tools compare them.
Put this guide to work
Know when the next run is missing
Connect one scheduled job, confirm a real completion, then verify the alert destination. Keep your existing scheduler in charge of running the work.