rp stock cpus

List reservable CPU instances with price and availability.

rp stock cpus [--dc <id>] [--vcpu N] [--product POD|SERVERLESS|CLUSTER] [--compact] [--json]

Options

  --dc <id>       keep only flavours stocked in this datacentre (case-insensitive)
  --vcpu N        keep only instances of N vCPUs (a power of two, minimum 2)
  --product <p>   POD, SERVERLESS or CLUSTER; show only that tier's price column
                  and drop flavours not offered on it (default: POD,SERVERLESS,
                  both columns)
  --compact       show one row per flavour (the old vcpu-range table) instead of
                  one row per power-of-two size
  --json          print the raw API response (the flavour objects, not expanded)

Notes

By default this verb EXPANDS each flavour into one row per deployable size: every power-of-two vCPU count inside the flavour's valid range (2, 4, 8, 16, 32, …). That is exactly the set rp pod create --cpu-flavor <ID> --vcpu <n> accepts, so each row is a real instance you can reserve. FLAVOUR is the flavour id rp pod create --cpu-flavor takes; VCPU and RAM_GB are that instance's size and total RAM (= VCPU × RAM_GB_VCPU). SECURE_PRICE and SERVERLESS_PRICE are the hourly rates for that instance size: the total (VCPU × per-vCPU rate) with the per-vCPU rate in parentheses. With --product the OTHER tier's column is dropped entirely, and any flavour not offered on that product (e.g. Memory-Optimized on SERVERLESS, whose serverless price is 0) is omitted from the table. A dash in the remaining price column means that tier is not offered. STOCK is shown only with --dc: it then reflects that datacentre's own availability (region-accurate). Without --dc the column is hidden, since the flavour-level value is only a global "available somewhere" aggregate. --vcpu N narrows the expanded rows to one size (validated: power of two,

= 2); --compact collapses back to one row per flavour (VCPU shown as a min-max range, RAM_GB_VCPU and the tier price as per-vCPU values). DATACENTERS lists every datacentre the flavour is stocked in, but the table column is truncated to the first two ids followed by "+N more" (e.g. "EU-CZ-1, EU-NL-1 +12 more") to keep the row readable; the full list is always in --json. Because filtering matches the underlying dataCenters array, --dc still finds a flavour even when its datacentre is hidden behind "+N more". With --dc the requested datacentre is omitted from the list (you already scoped to it). --dc and --product apply to BOTH the table and --json so the two views always show the same flavours.

API: GET /v2/catalog/cpus (include=AVAILABILITY, product=<selected>)