Cloud Instances

Provision compute capacity with explicit placement and health policies.

Choose an instance profile#

Instances provide compute capacity for managed services or dedicated workloads. A profile defines virtual CPUs, memory, and the supported storage class. Select capacity from observed workload demand rather than peak request counts alone.

Profile	vCPU	Memory	Typical use
general-2	2	4 GiB	Development and small APIs
general-4	4	8 GiB	Application services
memory-4	4	16 GiB	Caches and data processing

Declare placement#

apiVersion: universe/v1
kind: Instance
metadata:
  name: catalog-worker
spec:
  profile: general-4
  region: eu-central
  network: application
  bootVolumeGiB: 40
  labels:
    role: worker

Lifecycle and maintenance#

Instance states progress through pending, provisioning, ready, and draining. Planned maintenance drains service replicas before shutdown. A stopped instance retains its persistent volumes until an explicit deletion is approved.