Your cron jobs, background workers, and scheduled tasks run silently. When they fail — they fail silently too. Spectra gives each job a unique URL to call on completion. Miss the call, get the alert.
Heartbeat monitoring is the inverse of uptime monitoring. Instead of Spectra pinging you, your job pings Spectra. No ping within the expected window means something went wrong.
Name your job, set the schedule (interval or cron), and configure the grace period. Spectra generates a unique ping URL.
Add one line to the end of your script, cron job, or background task. Call the URL when the job completes successfully.
We expect a ping within your schedule + grace period window. A successful ping resets the clock.
If no ping arrives within the window, we immediately alert via Slack, email, Discord, webhooks, and more.
Append an optional ?status= parameter to the ping URL to signal job start, explicit failure, or to track run duration.
?status=start
Call the URL with ?status=start when your job begins, so we track the duration between start and complete — you know not just if it ran, but how long it took.
# Signal start
curl "spectrabolt.com/beat/...?status=start"
# ... job runs ...
curl "spectrabolt.com/beat/..." # complete
?status=complete
Calling the URL with ?status=complete resets the schedule window and records a successful run. Equivalent to a plain ping — use whichever is clearer in your code.
# Both are identical
curl "spectrabolt.com/beat/..."
curl "spectrabolt.com/beat/...?status=complete"
?status=fail
Call with ?status=fail when your job hits an error. This fires an immediate alert even if it ran within the window — for catching logical failures, not just missed runs.
# In your error handler:
curl "spectrabolt.com/beat/...?status=fail"
# Alert fires immediately
Copy the snippet for your language and paste it at the end of your script or job. Replace the URL with the one generated by Spectra for your monitor. Any HTTP method is accepted — POST, GET, or HEAD.
Any HTTP method accepted
POST, GET, or HEAD — all work equally. Use whichever is easiest in your stack. The URL itself identifies the monitor.
No authentication required
The unique UUID in the URL acts as the secret. Don't share it, but no auth headers are needed — just call the URL.
Don't ping on failure
Only ping the URL when your job succeeds. If the job errors out, don't call the URL (or call it with ?status=fail). Silence = alert.
Configure which failure conditions fire alerts per monitor. All selected conditions apply independently.
No ping received within the expected window + grace period. The silent failure.
Alert fires at: schedule + grace period
Example: daily + 5 min grace → alert at 24h 05m
The job triggered a call to the URL with the parameter which resulted in an immediate alert being generated.
Alert fires: immediately on ?status=fail call
No waiting: alert fires the moment Spectra receives it
Duration between ?status=start and complete exceeded the max.
Config: Maximum allowed duration
Example: set 30min max for an hourly job
Channels are reusable destinations — define them once in Settings → Alerts, then attach to individual monitors. Disabling a channel stops alerts for every monitor that uses it.
Answers to common questions teams ask us
Website monitoring checks something that should respond. Heartbeat monitoring watches for something that should call us — perfect for cron jobs and workers that have no public URL.
Any of them. If your job can make an HTTP request — curl, wget, a one-line library call — it can ping a Spectra heartbeat.
Extra time beyond the expected run before we consider a job late. It absorbs normal variance so you only get alerted on real problems.
Yes. Ping the /fail endpoint and Spectra raises an incident immediately, even if the job ran on schedule.
Yes — the job only needs outbound HTTP access to reach Spectra. No inbound connection to your infrastructure is required.
