What does a green check tell you about a daily report?
It tells you the check received a success signal. Send that signal only after the report reaches its intended destination and passes whatever checks matter for your app. Healthchecks.io checks the signal it receives. Your job still needs to check the report itself.
A heartbeat is a short message to the monitor. Keep the job’s private ping address out of public files and screenshots. Avoid attaching report contents or credentials to the message. The service can keep diagnostic text sent with a ping.
How long should the monitor wait before it alerts you?
Choose the schedule you expect and the extra time you can tolerate. That extra time is the grace period. If you also send a start signal, Healthchecks.io expects success within the grace period after that start. Set it long enough for a normal job to finish, and direct the alert to the person who can check the result.
What should you keep if you change monitors?
The hosted Hobbyist plan lists 20 jobs and 100 log entries per job. Keep your schedule, alert routes and job steps somewhere you control. A developer can use the Management API to read check settings and available logged signals. Reconnect the job to its new ping address and verify the new notification route when moving it.
The open-source Healthchecks software can also run on infrastructure you manage. You then own its updates, storage and notification delivery.
How can you tell a late report from a bad report?
Draw three schedules for a fictional report. Use paper only. Do not send a ping or connect a notification channel.
- On timeThe job saves a good sample report, then sends success. The monitor stays healthy.
- LateThe expected signal never arrives, even after the grace period. The monitor should route an alert.
- IncompleteThe sample report is empty. Your job should withhold success or report failure. The monitor cannot determine that from the file on its own.