Cron

Any cron entry can fire a heartbeat to AgentPing with one extra line for cost attribution and missed-run alerting.

The minimal example

0 9 * * *  /path/to/daily-summary.sh && /usr/bin/wget -qO- 'https://eu.ingest.agentping.io/v1/ping?key=ping_eu_018f3a2b9c1d7e8fa4b9c2d7e8f1a3b6&agent=daily-summary&status=ok' > /dev/null 2>&1

The trailing redirect keeps cron's email-on-output behaviour quiet. Drop it while debugging so you can see OK and any error text.

Reporting failures

Cron swallows the exit code of every job. To turn that into a signal, fire a different heartbeat on the failure path:

0 9 * * *  /path/to/daily-summary.sh && /usr/bin/wget -qO- 'https://eu.ingest.agentping.io/v1/ping?key=ping_eu_018f3a2b9c1d7e8fa4b9c2d7e8f1a3b6&agent=daily-summary&status=ok' > /dev/null 2>&1 || /usr/bin/wget -qO- 'https://eu.ingest.agentping.io/v1/ping?key=ping_eu_018f3a2b9c1d7e8fa4b9c2d7e8f1a3b6&agent=daily-summary&status=fail' > /dev/null 2>&1

&& runs the success ping only if the script exited zero; || runs the failure ping only if it didn't. Both URLs target the same agent.

For long jobs, fire start at the top and ok or fail at the bottom; the server computes duration from the gap. See Heartbeats.

Schedule monitoring

A heartbeat that never fires is invisible by default. Set the agent's expected cron in the dashboard and Pulse pages you on missed runs. Use the same cron syntax in both places.

Header-auth variant

For environments that log every URL through a proxy:

0 9 * * *  /usr/bin/curl -fsS -H "X-AgentPing-Key: ping_eu_018f3a2b9c1d7e8fa4b9c2d7e8f1a3b6" 'https://eu.ingest.agentping.io/v1/ping?agent=daily-summary&status=ok' > /dev/null 2>&1

curl -fsS is the cron-friendly invocation: fail silently on connection errors, show the body on HTTP errors, no progress meter.