Unexpected charge: 40 idle machines left running on pl-factory-worker

Questions / Help
1

Hi,

I was charged around $600 for my app pl-factory-worker (org: Personal, region: NRT).

The app had 41 machines running (performance-1x-cpu@2048MB), in two process
groups: app (40) and gsp (1). My fly.toml only declares a single VM — the
extra machines were created manually during a batch of parallel compute runs and
were never scaled back down. They then sat idle for roughly two weeks before I
noticed, because the only workstation with flyctl installed had been retired, so
I had no way to see or stop them.

I have already stopped 39 of them and kept 2 running.

I understand the usage is real and this was my mistake. I’m asking whether you
can apply any credit for the idle period, since none of that compute did any
work. Even a partial credit would help a lot.

I’d also like to set up a usage/spending alert so this can’t repeat — could you
point me at the right setting?

Thanks,

2

Hello 카피더밸류! You can share this with our billing team here: billing@fly.io.

Our Billing team’s got you: they can help you on this request, and point you on the right direction

3

It’s worth mentioning Accident Forgiveness, if you’d like to claim on that. See also: