Command 'flyctl platform vm-sizes --json' does not print output in JSON format
flyctl v0.3.56
Command output:
Machines platform
NAME CPU CORES MEMORY
shared-cpu-1x 1 256 MB
shared-cpu-2x 2 512 MB
shared-cpu-4x 4 1 GB
shared-cpu-8x 8 2 GB
NAME CPU CORES MEMORY
performance-1x 1 2 GB
performance-2x 2 4 GB
performance-4x 4 8 GB
performance-8x 8 16 GB
performance-16x 16 32 GB
NAME CPU CORES MEMORY GPU MODEL
a10 8 32 GB a10
a100-40gb 8 32 GB a100-pcie-40gb
a100-80gb 8 32 GB a100-sxm4-80gb
l40s 8 32 GB l40s
Hi @gskierk, I would like to contribute to fixing this bug. Can I get any pointers? Like how output should look.
Regards Mahathi
Hi @gskierk, I would like to contribute to fixing this bug. Can I get any pointers? Like how output should look.
Regards Mahathi
I don't know how it looked before this regression, sorry. Also, there is no structure of the data in the documentation here: https://fly.io/docs/flyctl/platform-vm-sizes/
But maybe fixing it will just restore the old structure.
I wasn't able to find a reference on how it used to look, so I made it look like the following:
{
"shared": [
{
"name": "shared-cpu-1x",
"cpu_cores": "1",
"memory": "256 MB"
},
{
"name": "shared-cpu-2x",
"cpu_cores": "2",
"memory": "512 MB"
},
{
"name": "shared-cpu-4x",
"cpu_cores": "4",
"memory": "1 GB"
},
{
"name": "shared-cpu-8x",
"cpu_cores": "8",
"memory": "2 GB"
}
],
"performance": [
{
"name": "performance-1x",
"cpu_cores": "1",
"memory": "2 GB"
},
{
"name": "performance-2x",
"cpu_cores": "2",
"memory": "4 GB"
},
{
"name": "performance-4x",
"cpu_cores": "4",
"memory": "8 GB"
},
{
"name": "performance-8x",
"cpu_cores": "8",
"memory": "16 GB"
},
{
"name": "performance-16x",
"cpu_cores": "16",
"memory": "32 GB"
}
],
"gpu": [
{
"name": "a10",
"cpu_cores": "8",
"memory": "32 GB",
"gpu_model": "a10"
},
{
"name": "a100-40gb",
"cpu_cores": "8",
"memory": "32 GB",
"gpu_model": "a100-pcie-40gb"
},
{
"name": "a100-80gb",
"cpu_cores": "8",
"memory": "32 GB",
"gpu_model": "a100-sxm4-80gb"
},
{
"name": "l40s",
"cpu_cores": "8",
"memory": "32 GB",
"gpu_model": "l40s"
}
]
}
If this looks okay to the Fly.io team, I can open a pull request.