As part of T342750: cloud: introduce a kubernetes undercloud to run openstack (via openstack-helm) we will be deploying a new ceph cluster in the codfw datacenter.
Before implementation, we may want to decide on some details, like initial minimum capacity, rack distribution, etc.
As a first approach, we might want to start with 8 ceph nodes, spread in 4
racks for high availability (we have 3 replicas). Any less nodes and the high
availability gets compromised and we risk the cluster halting if a rack goes
down (and with one node per rack, a single node going down makes the rack go
down xd).