Run commands and retry offline reports
Report a command's execution while preserving its output and exit status.
On this page
What you need
A Job or Workflow monitor, a configured CLI identity and a device key.
How to do it
- Run metribeacon run --monitor-id MONITOR_UUID -- PROGRAM ARGUMENTS. The command after -- is executed directly.
- Use --strict only if reporting failure should also make an otherwise successful command fail. Use --scheduled-for when a scheduled run needs an explicit target time.
- After a connection problem, run metribeacon run flush to send buffered reports. Check the run history before executing the real task again.
Good to know
The wrapper does not add a shell. It preserves the command's input, output and exit or signal status. In strict mode, a successful command with unsent reports can return 75.
The offline buffer survives restarts but has size and age limits. A failed report is separate from a failed task.
metribeacon run --monitor-id MONITOR_UUID --strict -- /usr/local/bin/demo-job
metribeacon run flush --jsonIf something goes wrong
Separate the task's error from a reporting error. Flushing sends saved reports; it does not execute the task again.
Related guides
- Monitor jobs and cron schedules
Track scheduled commands and find failed, late or missing executions.
- Workflows and parallel runs
Follow each workflow or AI-agent execution, including simultaneous runs.
- Complete command reference
Find the right command and use structured output for automation.