Regions & stock

How Cloud API decides where to place a server across every instance and node.

Region availability

GET /v1/regions summarizes deployable capacity per region:

json
{
  "regions": [
    { "region": "USA",   "freeMemory": 131072, "freeSlots": 48, "unlimited": false, "deployable": true },
    { "region": "SPAIN", "freeMemory": 65536,  "freeSlots": 20, "unlimited": false, "deployable": true }
  ]
}

A region is deployable when at least one live (non-maintenance) node has a free allocation and ≥256 MB free memory. unlimited is true when a node has no capacity cap configured.

The stock engine

The engine mirrors the panel's viability rules:

A node limit of 0 means unlimited for that dimension.
An overallocation of -1 also means unlimited.
Otherwise effective capacity = limit × (1 + overallocation%).
When a node sets no overallocation, the platform default (50%) applies.
Usage is computed from the real server build limits on each node.

Placement

When you deploy, Cloud API filters to nodes that:

are not in maintenance,
have enough free memory, disk and CPU for the request,
have at least one free allocation,
match the requested region (if any).

Candidates are sorted by instance priority, then by most free memory. Passing region: null lets the engine choose the best region globally.