Upstate
Scheduled job and backup check-ins
Check-ins answer whether a job you operate reported on schedule. They are useful for backups, cleanup tasks, recurring jobs and other work that should finish regularly. Upstate does not run the job or verify its output.
Create a check-in
Open Check-ins. Give the job a name and kind (Backup, Scheduled job or Other), then set Expected every and Grace period in seconds. The form offers optional Project and Server selectors; leave them as Account-wide and No server if the job is not tied to either. There is no environment selector on this screen. Upstate shows a private ping URL when you create the check-in; copy it into your own job. Never publish that URL, commit it or paste it into a public incident report.
The generated command uses a bounded HTTP request with retries. Follow the exact command on the Check-ins page rather than copying another account's token.
Report a result
A plain ping records a successful finish. The generated URL can also receive a /start suffix at the beginning of a job and a /fail suffix (or a nonzero exit-code suffix) when the job fails. Use the URLs and exit-code handling shown in your account so a script that fails does not accidentally send a success ping.
A start without a matching completion may become Stuck. A never-run check-in is Waiting for first ping, not healthy; once its interval and grace pass, the status becomes late or down. A reported failure can alert immediately. Pausing a check-in prevents missed-job alerts while the job is intentionally offline.
Interpret check-in status
| Status | Interpretation |
|---|---|
| Waiting for first ping | The check-in exists but no finish has been reported yet. |
| Late | Its due time passed and the grace window is still open. |
| Down | It is past the expected interval and grace without a finish. |
| Failed | The job explicitly reported failure. |
| Stuck | The job reported a start but no finish within the allowed window. |
| Paused | Scheduling alerts are intentionally disabled. |
An account owner/admin can manage, rotate, pause and delete check-ins. Rotating the token immediately invalidates the previous URL. Check-ins and their alert history are account-scoped.
Backup check-ins are not backup verification
A successful ping proves only that your script sent success. It does not prove a completed backup is consistent, stored offsite or restorable. Keep independent retention checks, error handling and restore drills. Production Audit may flag a recognized database accessory that lacks a reporting backup check-in.
Route misses or failures through Alerts and notifications, and see Production audit for evidence classifications.