Container Service for Kubernetes:Specify the compute specification when creating an elastic container instance-based pod

Example 1: Specify GPU instance types

Use GPU instance types and set the nvidia.com/gpu resource limit per container. GPUs are shared across containers in the pod.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: test
  labels:
    app: test
spec:
  replicas: 2
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      name: nginx-test
      labels:
        app: nginx
        alibabacloud.com/eci: "true"
      annotations:
        k8s.aliyun.com/eci-use-specs: "ecs.gn6i-c4g1.xlarge,ecs.gn6i-c8g1.2xlarge" # Specify a maximum of five GPU-accelerated ECS instance types at a time.
    spec:
      containers:
      - name: nginx
        image: registry.cn-shanghai.aliyuncs.com/eci_open/nginx:1.14.2
        resources:
            limits:
              nvidia.com/gpu: "1" # The number of GPUs required by the Nginx container. The GPUs are shared.
        ports:
        - containerPort: 80
      - name: busybox
        image: registry.cn-shanghai.aliyuncs.com/eci_open/busybox:1.30
        command: ["sleep"]
        args: ["999999"]
        resources:
            limits:
              nvidia.com/gpu: "1" # The number of GPUs required by the BusyBox container. The GPUs are shared.

Example 2: Mix vCPU/memory specs and ECS instance types

List specs in priority order. ECI tries them left to right until one succeeds.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: test
  labels:
    app: test
spec:
  replicas: 2
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      name: nginx-test
      labels:
        app: nginx
        alibabacloud.com/eci: "true"
      annotations:
        k8s.aliyun.com/eci-use-specs: 2-4Gi,ecs.c5.large,ecs.c6.large  # Sets the specifications that you want to use to create the pod. Replace the value by using the actual specifications.
    spec:
      containers:
      - name: nginx
        image: registry.cn-shanghai.aliyuncs.com/eci_open/nginx:1.14.2
        ports:
        - containerPort: 80

Verify the spec used

After the pod starts, check the k8s.aliyun.com/eci-instance-spec field in the YAML file of the pod to confirm which specification ECI actually allocated and verify your billing basis.

vCPU and memory specifications

Preemptible instances

Preemptible instances can reduce costs significantly for stateless and fault-tolerant workloads. See Create a preemptible elastic container instance.

Reserved instances and savings plans

For long-running stable workloads, offset your ECI bills with reserved instances or savings plans:

• vCPU/memory-billed pods: Use general-purpose savings plans.

• ECS instance type-billed pods: Use general-purpose savings plans, ECS compute savings plans, or reserved instances.

See Use reserved instances and Use savings plans.