Grouping of cloud provider instance types/machine types/VM sizes list in cluster creation

[marians]

User story

• As an admin creating a cluster or node pool, I want to make an informed decision when selecting a machine type to use for nodes.

Details

Currently we offer all available instance types / machine types / VM sizes in a flat list/table. The large number of types to choose from makes this list hard to use.

All cloud providers we deal with currently have a notion of a category, family, or group each type belongs to. Representing this grouping in our selection widget would simplify the selection.

Provider taxonomies

Google Cloud differentiates between "machine families", which fall into the following schema:

• General-purpose workloads
  • Cost-optimized
  • E2
  • Balanced
  • N2, N2D, N1
  • Scale-out optimized
  • Tau T2D, Tau T2A
• Optimized workloads
  • Memory-optimized
  • M3, M2, M1
  • Compute-optimized
  • C2, C2D
  • Accelerator-optimized
  • A2

Azure presents the following system:

• General purpose
  • B, Dsv3, Dv3, Dasv4, Dav4, DSv2, Dv2, Av2, DC, DCv2, Dpdsv5, Dpldsv5, Dpsv5, Dplsv5, Dv4, Dsv4, Ddv4, Ddsv4, Dv5, Dsv5, Ddv5, Ddsv5, Dasv5, Dadsv5
• Compute optimized
  • F, Fs, Fsv2, FX
• Memory optimized
  • Esv3, Ev3, Easv4, Eav4, Epdsv5, Epsv5, Ev4, Esv4, Edv4, Edsv4, Ev5, Esv5, Edv5, Edsv5, Easv5, Eadsv5, Mv2, M, DSv2, Dv2
• Storage optimized
  • Lsv2, Lsv3, Lasv3
• GPU
  • NC, NCv2, NCv3, NCasT4_v3, ND, NDv2, NV, NVv3, NVv4, NDasrA100_v4, NDm_A100_v4
• High performance compute
  • HB, HBv2, HBv3, HC

AWS EC2:

• General purpose
  • M7g, Mac, M6g, M6i, M6in, M6a, M5, M5n, M5zn, M5a, M4, A1, T4g, T3, T3a, T2
• Compute optimized
  • C7g, C7gn, C6i, C6in, C6a, C6g, C6gn, C5, C5n, C5a, C4
• Memory optimized
  • R7g, R7iz, R..., X2gd, X2iedn, X..., High Memory, z1d
• Accalerated computing
  • P..., DL1, Trn1, Inf..., G5, G5g, G4..., G3, F1, VT1
• Storage optimized
  • Im4gn, Is4gen, I4i, I3, I3en, D2, D3, D3en, H1
• HPC optimized
  • Hp6id, Hpc6a

[marians]

Keeping this, as it's relevant once we port cluster creation into Backstage.