Configuring QoS for a Pod

Scenario

Bandwidth preemption occurs between different containers deployed on the same node, which may cause service jitter. You can configure bandwidth limitation for the pod to solve this problem.

Specifications

Using the CCE Console

When creating a workload on the console, you can set pod ingress and egress bandwidth limits by clicking Network Configuration in the Advanced Settings area.

Using kubectl

apiVersion: apps/v1
kind: Deployment
metadata:
  name: test
  namespace: default
  labels:
    app: test
spec:
  replicas: 2
  selector:
    matchLabels: 
      app: test
  template:
    metadata:
      labels:
        app: test
      annotations:
        kubernetes.io/ingress-bandwidth: 100M
        kubernetes.io/egress-bandwidth: 100M
    spec:
      containers:
        - name: container-1
          image: nginx:alpine
          imagePullPolicy: IfNotPresent
      imagePullSecrets:
        - name: default-secret

• kubernetes.io/ingress-bandwidth: ingress bandwidth of the pod
• kubernetes.io/egress-bandwidth: egress bandwidth of the pod

If these two parameters are not specified, the bandwidth is not limited.

After modifying the ingress or egress bandwidth limit of a pod, restart the container for the modification to take effect. After annotations are modified in a pod not managed by workloads, the container will not be restarted, so the bandwidth limits do not take effect. You can create a pod again or manually restart the container.