Docs › Fly.io scheduled Machines with RunVouch

Fly.io scheduled Machines with RunVouch

On Fly you either run a Machine with a schedule (fly machine run --schedule daily) or run cron (often supercronic) inside a long-lived Machine. Both stop silently when the Machine is destroyed, stuck or out of memory.

How it runs on Fly.io scheduled Machines

The Machine's command, or the crontab line inside the container. The image needs Python 3 for rv.

Store the key

fly secrets set RUNVOUCH_KEY=rv_...; secrets are exposed as environment variables in every Machine of the app.

Wrap the job

# scheduled Machine (runs once per schedule, then stops)
fly machine run ghcr.io/you/report:latest --schedule daily \
  --command "rv run nightly-report --log /tmp/report.log -- python report.py"

# or, inside a long-lived Machine with supercronic
0 2 * * * rv run nightly-report --evidence-file /data/report.html -- python /app/report.py

rv fails open: if RunVouch is unreachable the job still runs and you get one warning line.

Register the cadence and caps

rv agent nightly-report --cadence 24h --grace 30m --max-runtime 1h --evidence --cap-run-cost 2

Register the agent once, from anywhere with the key. Cadence is what turns a schedule that stopped into an alert; --evidence makes a run without evidence a failure; the caps pause the agent when it overspends.

What goes silent on Fly.io scheduled Machines

• Scheduled Machines run "approximately" at the interval; use a grace of an hour or more.
• An OOM-killed Machine restarts (or does not); the run that was in flight never ends and becomes STALLED.
• A Machine that was stopped by fly scale count 0 or a failed deploy takes the cron with it; MISSED is the signal.

What RunVouch detects

Need a key? Get a free key · Stuck? contact