I spin up extra machines to deploy migrations either on a release:
[deploy]
release_command = "npx prisma migrate deploy"
or when I load data from prod. I run the following script in a terminal:
fly console -a {service_name} -C "npx prisma migrate deploy
Sometimes the script is stopped or it fails. So what I have observed is that the machines are not closed down.
Here I have uploaded a screen shot where you can see a console machine that has been for 10 days. So would like to know how to automatically clean these up. Either by running the commands differently or by given a script that can kill all machines running fly_app.console
which I can set up to run every night as a cron job.