As outlined in T239151#5707691 the creation of a single 16G Ganeti VM in eqiad row_A currently fails with...
Failure: prerequisites not met for this operation: error type: insufficient_resources, error details: Can't compute nodes using iallocator 'hail': Request failed: Group row_A (preferred): No valid allocation solutions, failure reasons: FailMem: 12
Which indicates that we've run low on available memory in the eqiad row_A ganeti group.
The Eqiad Ganeti expansion being tracked in T228924#5786002 may help to some degree, in that it will allow us to move VMs from the row_A group to row_B (or row_D, depending on the TBD final racking location of the hardware).
Still, IMO we should expand capacity on eqiad row_A if we have the resources available to do so.
The current row_A hosts look to be configured with 12 DIMM slots, 4 of which are populated with 16G sticks. So there is physical capacity to e.g. double row_A memory by adding (16) 16G sticks of ram (4 sticks per host, across 4 hosts)
But since the current hardware is from 2017 and support expires this April, we probably should expand the row_A group with new servers. Refreshing with a set of 4 new 128G RAM nodes would double memory capacity in the current rack (and roughly power) footprint (after decom of the 2017 row_A hardware)