Skip to content

Latest commit

 

History

History
38 lines (38 loc) · 1.11 KB

limits.md

File metadata and controls

38 lines (38 loc) · 1.11 KB

resource requests

  • k8s will intelligently schedule to nodes based on the resource requirements
  • by default the cpu and memory are limited to 1 and 512Mi respectively. ^24a178
  • if no [[node]] with sufficient resources are available, [[pods]] will be pending state
  • cpu values can be as low as 0.1 or 1m (m stands for milli)
  • 1 count equals to 1 vCPU in AWS or 1 core in GCP and Azure or 1 Hyperthread
  • memory is denoted in either G/M/K (1000) or Gi/Mi/Ki (1024).
  • limits section in pod definition can be used set new [[#^24a178|default limits]]
  • for the pod to pick up those defaults you must have first set those as default values for request and limit by creating a LimitRange in that namespace.

definition

---
apiVersion: v1
kind: LimitRange
metadata:
  name: mem-limit-range
spec:
 resources:
   limits:
    - default:
       memory: "512Mi"
      defaultRequest:
       memory: "256Mi"
      type: Container
---
apiVersion: v1
kind: LimitRange
metadata:
  name: cpu-limit-range
spec:
 resources:
   limits:
    - default:
       cpu: 1
      defaultRequest:
       cpu: 0.5
      type: Container