I have set up a grafana instance which is used rarely such that it is configured to scale to zero. The delay in response if needed is OK.
At the same time, the instance has grafana alerts configured and obviously, this only works if the machine runs from time to time. So, I have configured the machine to run every hour using fly machine update … --schedule hourly. That works fine, the machine is started hourly and after a while gets scaled to zero because excess capacity is detected.
But, if I now try to use the machine in between, the machine fails to get spun up and I see these errors in the log:
2026-02-23 17:27:16.527 [PM07] failed to change machine state: machine still active, refusing to start
2026-02-23 17:27:16.526 Starting machine
2026-02-23 17:27:13.521 [PM07] failed to change machine state: machine still active, refusing to start
2026-02-23 17:27:13.517 Starting machine
2026-02-23 17:27:12.332 [PM07] failed to change machine state: machine still active, refusing to start
2026-02-23 17:27:12.331 Starting machine
2026-02-23 17:27:12.157 [PM07] failed to change machine state: machine still active, refusing to start
2026-02-23 17:27:12.156 Starting machine
However, I can manually fly machine start just fine. Is there a way to fix this to get both parts working?