In this short article, I’ll go through the resource limits and quota in Kubernetes.

LimitRange

LimitRange is an API object that limits resource usage per container or Pod in a Namespace:

A LimitRange is a Kubernetes object used to:

  • Automatically apply default CPU/memory limits and requests to pods/containers in a namespace

  • Prevent pods from over-consuming resources

  • Avoid the need to define limits in every pod manually

  • works only within a single namespace.

Key Fields in LimitRange:

  1. type: Defines the scope of the limit:

    • Container: applies to individual containers

    • Pod: applies to the entire pod

    • PersistentVolumeClaim: applies to storage requests

  2. defaultRequest:

    • If a container does not specify a request, this value is used automatically.

    • Request = the amount of resource guaranteed for the container.

  3. default:

    • If a container does not specify a limit, this value is used automatically.

    • Limit = the maximum amount the container is allowed to use.

Example:

apiVersion: v1
kind: LimitRange
metadata:
  name: mem-cpu-limitrange
  namespace: dev
spec:
  limits:
  - type: Container
    default:
      cpu: "500m"
      memory: "512Mi"
    defaultRequest:
      cpu: "250m"
      memory: "256Mi"

What this does:

  • If a container has no resource settings, it will:

    • Request: 250m CPU, 256Mi memory. It reservse 250m CPU & 256Mi memory for the container

    • Max Limit: 500m CPU, 512Mi memory. It restricts it to max 500m CPU & 512Mi memory

Quota

A ResourceQuota is a Kubernetes API object that:

  • Limits the total amount of resources (CPU, memory, storage, objects) that can be used in a namespace.

  • Helps control resource usage at the namespace level, preventing a single team or app from using everything.

Some difference from LimitRange:

LimitRange ResourceQuota
Applies per container/pod Applies to the entire namespace
Sets default values or max values per pod Sets overall cap for the namespace
Does not require the pod to specify resource requests Requires pods to specify resources.requests and/or resources.limits

Example ResourceQuota declarion:

apiVersion: v1
kind: ResourceQuota
metadata:
  name: compute-resources
  namespace: dev
spec:
  hard:
    requests.cpu: "2"
    requests.memory: "2Gi"
    limits.cpu: "4"
    limits.memory: "4Gi"

We can create ResourceQuota using cli.

kubectl create quota my-quota --hard cpu=100m,memory=500Mi,pods=3 -n limited

you can get a create command using kubectl create quota --help

To describe the created quota:

kubectl describe quota -n limited

Once the deployment created I noticed that there were no pods were started. After quick analyze I found out that the reason was that I didn’t set any resource restriction.

kubectl -n limited describe replicasets.apps nginx-5869d7778c

To solve the issue, I needed to apply resource to the deployment:

kubectl -n limited set resources deploy nginx --requests cpu=100m,memory=5Mi --limits cpu=200m,memory=20Mi

you can get a create command using kubectl set resources --help

To change the current quota we should edit it using:

kubectl -n limited edit quota my-quota

Restart deployment:

kubectl -n limited rollout restart deployment nginx

You may also enjoy