Resource Optimization

Purpose: For platform engineers, provides guidance on right-sizing workloads, configuring autoscaling, and managing resource quotas across multi-tenant clusters.

Right-Sizing Worker Pools

Use dedicated node pools for different workload profiles:

Pool	Instance Profile	Use Case
system	4 vCPU / 16 GB	Platform services (observability, security, GitOps)
general	8 vCPU / 32 GB	Stateless application workloads
memory	4 vCPU / 64 GB	Caches, in-memory databases
compute	16 vCPU / 32 GB	Batch processing, CI runners

Horizontal Pod Autoscaling (HPA)

openCenter clusters ship with metrics-server pre-installed. Configure HPA using standard Kubernetes resources:

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: app-hpa
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: my-app
  minReplicas: 2
  maxReplicas: 20
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 70

Vertical Pod Autoscaling (VPA)

Deploy VPA in recommendation mode to identify right-sizing opportunities without automatic mutation:

updatePolicy:
  updateMode: "Off"  # Recommendations only

Resource Quotas for Multi-Tenancy

Enforce per-namespace quotas to prevent noisy-neighbor problems:

Resource	Recommended Default	Notes
requests.cpu	8 cores	Per namespace
requests.memory	32 Gi	Per namespace
limits.cpu	16 cores	2× requests for burst
persistentvolumeclaims	20	Prevent storage exhaustion
services.loadbalancers	2	MetalLB IP conservation

LimitRange Best Practices

Set default requests/limits so pods without resource specs don't consume unbounded resources:

apiVersion: v1
kind: LimitRange
metadata:
  name: default-limits
spec:
  limits:
  - default:
      cpu: 500m
      memory: 512Mi
    defaultRequest:
      cpu: 100m
      memory: 128Mi
    type: Container