GrpcBackendGroup resource fields

GrpcBackendGroup enables you to group backends, i.e., Kubernetes services processing gRPC traffic. The Application Load Balancer Ingress controller uses these resources to create backend groups.

You need to add a reference to GrpcBackendGroup to the Ingress resource. This Ingress resource must have the ingress.alb.yc.io/protocol: grpc annotation.

---
apiVersion: alb.yc.io/v1alpha1
kind: GrpcBackendGroup
metadata:
  namespace: testapp-ns
  name: grpc-bg
spec:
  backends:
    - name: ying
      weight: 50
      service:
        name: grpc-testapp-service
        port:
          number: 8080

    - name: yang
      weight: 50
      service:
        name: grpc-testapp-service2
        port:
          number: 8080
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: testapp-ingress-with-grpc-bg
  namespace: testapp-ns
  annotations:
    ingress.alb.yc.io/group-name: default
    ingress.alb.yc.io/protocol: grpc
spec:
  rules:
    - host: grpc-first-server.info
      http:
        paths:
          - path: /helloworld.Greeter/SayHello
            pathType: Prefix
            backend:
              resource:
                apiGroup: alb.yc.io
                kind: GrpcBackendGroup
                name: grpc-bg

GrpcBackendGroup is a custom resource from the alb.yc.io API group provided by an Ingress controller.

GrpcBackendGroupGrpcBackendGroup

apiVersion: alb.yc.io/v1alpha1
kind: GrpcBackendGroup
metadata:
  name: <string>
spec:
  backends:
    - name: <string>
      weight: <int64>
      service:
         name: <int64>
         port:
           name: <string>
           number: <int32>
      tls:
        sni: <string>
        trustedCa: <string>
      healthChecks:
        - grpc:
            serviceName: <string>
          port: <int32>
          healthyThreshold: <int32>
          unhealthyThreshold: <int32>
          timeout: <string>
          interval: <string>
    - ...

Where:

• apiVersion: alb.yc.io/v1alpha1

• kind: GrpcBackendGroup

• metadata (ObjectMeta; this is a required field)

  Resource metadata.

  • name (string; this is a required field)

    Resource name. For more information about the format, see this Kubernetes guide.

    You must specify this name in the spec.rules.http.paths.backend.resource.name field of the Ingress resource (see this configuration).

    Do not mistake this name for the Application Load Balancer backend group name.

• spec (GrpcBackendGroupSpec)

  Resource specification.

  • backends ([]GrpcBackend)

    List of backends in the group.

    • name (string; this is a required field)

      Backend name.

    • weight (int64)

      Backend weight. Backends in a group receive traffic in proportion to their weights.

      You should either specify weights for all backends in a group, or not specify them at all. If weights are not specified, traffic will be equally distributed across backends.

      A backend with zero or negative weight will not be receiving traffic.

    • service (ServiceBackend)

      Reference to the Kubernetes service to process requests as a backend.

      The referred Service resource must be described per the standard configuration.

      • name (string, required):

        Kubernetes service name.

      • port (ServiceBackendPort, required):

        Service port to which Ingress will direct traffic.

        The field is designed for ingress controller operation and has no equivalents in Application Load Balancer.

        • name (string):

          Service port name.

          This name must match one of the Service resource spec.ports.name values. For more information, see the resource specification.

          You must specify either the service port name or its number, but not both.

        • number (int32):

          Service port number.

          This number must match one of the Service resource spec.ports.port values. For more information, see the resource specification.

          You must specify either the service port name or its number, but not both.

    • tls (BackendTLS)

      TLS connection settings for the load balancer nodes and backend endpoints.

      If this field is specified, the load balancer will establish TLS connections to the backend, comparing received certificates with the one specified in the trustedCa field. Otherwise, the load balancer will use unencrypted connections to the backend.

      • sni (string)

        SNI domain name for TLS connections.

      • trustedCa (string)

        Contents of the X.509 certificate issued by a certificate authority in PEM format.

    • healthChecks ([]HealthChecks)

      Custom health checks settings for Managed Service for Kubernetes cluster applications.

      By default, the Application Load Balancer Ingress controller receives L7 load balancer health check requests on TCP port 10501. Then it checks kube-proxy pods on each cluster node. Given that kube-proxy is healthy, the process is as follows: if an application does not respond in a particular pod, Kubernetes redirects traffic to a different pod or node.

      You can use healthChecks settings to customize application health checks.

      • grpc (GrpcBackend)

        Specifies gRPC as the health check protocol.

        • serviceName (string)

          Name of the service you want to check.

      • port (int32)

        Cluster node port for checking application availability. This port should match the spec.ports.nodePort value of the NodePort Service resource.

        The application will be available for health checks at http://<node_IP_address>:<port>/<path>.

      • healthyThreshold (int32)

        Number of consecutive successful checks required to consider the application endpoint healthy.

      • unhealthyThreshold (int32)

        Number of consecutive failed checks required to consider the application endpoint unhealthy.

      • timeout (string)

        Response timeout in seconds. You can specify values between 1s and 60s.

      • interval (string)

        Health check request interval in seconds.

        You can specify values between 1s and 60s. interval must exceed timeout by at least one second.

      Note

      You can also configure application health checks using the ingress.alb.yc.io/health-checks annotation of the Service resource.