Your cron job died quietly. Find out in minutes, not weeks.
Add one curl line to the end of any job. Alertee expects the ping on your schedule and alerts you the moment it stops arriving — or the moment the number it reports goes wrong. No agent, no credentials, no inbound access to anything you run.
# last line of your backup / cron / sync script
curl -fsS "$ALERTEE_PING_URL"
# or push a number and alert on its value too
curl -X POST "$ALERTEE_INGEST_URL" \
-d '{"value": 1842003}' # rows copied tonightYou get your unique URL the moment you create the check.
Free plan, no card · Works from behind NAT · Email, Slack, webhook · 2-minute setup
The failures that never page anyone
Uptime monitors watch things that answer. A dead man's switch watches things that are supposed to speak up on their own — and tells you when they go silent.
Backups that stopped backing up
The classic. The backup script broke months ago and nobody knew until restore day. Ping after every successful run — or push the byte count and alert when it collapses.
Cron jobs that die quietly
Cron doesn't email anyone anymore. A job that stops running produces no error, no log line, nothing — just an absence. A dead man's switch turns that absence into an alert.
Anything that can run curl
Cert renewals, sync scripts, RAID health checks, sensor readings, queue drains. If it runs on a schedule and can hit a URL, it can be watched — homelab or production.
How it works
Two check types, one idea: your job reports in, and silence is the alarm.
Create a check, get a URL
Sign up, pick 'heartbeat' or 'push', set how often the job should report in. You get a unique URL immediately — the whole thing takes about two minutes, phone included.
Add one line to your job
A curl at the end of the script, after the real work succeeds. That's the entire integration. No agent to install, no library, no config file.
Silence becomes an alert
If the ping doesn't arrive inside the window, Alertee opens an incident and alerts you by email, Slack, or webhook. When the job recovers, the incident resolves.
Heartbeat or push — which one?
A heartbeat check is the pure dead man's switch: ping the URL, silence triggers the alert. A push check does that and carries a value — rows copied, bytes written, items in the queue — so you also get alerted when the job runs fine but the number is wrong. Both are one curl. Start with heartbeat, upgrade the line to push when you want the number watched too.
Nothing to install, nothing to expose
No credentials
Alertee never logs into your servers or databases for this. Your job pushes one request out; we never reach in.
Works from anywhere
Homelab behind NAT, air-gapped-ish VPC, a Raspberry Pi in the loft. Outbound HTTPS is the only requirement.
No agent, no library
The integration is a URL. curl, wget, an HTTP call from any language, a CI step — whatever your job already has.
Five switches, free, forever
Create a check, add the curl line, and stop wondering whether last night's backup actually ran. No credit card required.
5 checks, 2 data sources. No card.
25 checks, 5 data sources, 5-min intervals.
500 checks, 50 data sources, 1-min intervals.