When doing a deploy, I’m getting an OOM on the release machine
The machine size is shared-cpu-1x@256MB
The docs say:
" The temporary Machine inherits the size from the largest Machine in the default process group of the app as of flyctl v0.0.508 (they also default larger on empty/new apps, using the Machine shared-cpu-2x preset)."
our machines are mostly shared-cpu-4x@1024MB, however the machines in our beat process group (for redis worker beats) are shared-cpu-1x@256MB
It seems like the release machine is scaling to that size. I’m going to try scaling the beat machine(s) but I wonder if there’s a way to know what the “default process group” is?
after scaling the beat process group to shared-cpu-2x@1024, the release machine now matches that at shared-cpu-2x@1024MB, and the deploy succeeded.
I think the issue here is adopting the machine size of the default group doesn’t give the user enough control over the size of the deployment machine, and it’s also unclear what the default process group is.
Agreed… The process groups docs say that it is app in one fallback case; the source code itself describes something fairly complicated:
// DefaultProcessName returns:
// * "app" when no processes are defined
// * "app" if present in the processes map
// * The first process name in ascending lexicographical order
It sounds like your own names might be beats and web—or something along those lines.
You might be able to work around this by just having an artificial app group that really means “release command Machines” to you, .
.