Heartbeat checks

How it works

1. Add a check

   Give it a name, how often the job runs (from every minute to once a week), how late it may be, and which app sends the alert.

2. Ping it from the job

   Call the check's URL at the end of the job, only when the job succeeded. The check starts with the first ping, and never alerts before it.

3. Hear about the silence

   If no ping arrives within the period plus the time it may be late, you get a high-priority "Name is down" on all the app's devices, and "Name is back up" with the next ping.

Examples

Cron

0 3 * * * /usr/local/bin/backup.sh && curl -fsS -m 10 --retry 3 https://api.bzzd.io/ping/PING_KEY > /dev/null

At the end of a shell script

curl -fsS -m 10 --retry 3 https://api.bzzd.io/ping/PING_KEY

Without curl: wget -q -T 10 -O /dev/null https://api.bzzd.io/ping/PING_KEY

Windows (PowerShell, Task Scheduler)

Invoke-RestMethod -TimeoutSec 10 https://api.bzzd.io/ping/PING_KEY

Any tool that can open a URL works: GET, HEAD and POST all count, and the answer is a plain OK. A ping never sends a notification by itself, and doesn't count towards your messages; only the alerts do.

Details

• Pause a check while a job is meant to be off, like during maintenance; it ignores pings and never alerts while paused.
• Changing how often a check runs keeps its URL, so nothing needs updating on the server.
• Free includes 5 heartbeat checks.
• Pro includes 50 heartbeat checks.