Workload Group

4 minute read

WorkloadGroup describes a collection of workload instances. It provides a specification that the workload instances can use to bootstrap their proxies, including the metadata and identity. It is only intended to be used with non-k8s workloads like Virtual Machines, and is meant to mimic the existing sidecar injection and deployment specification model used for Kubernetes workloads to bootstrap Istio proxies.

The following example declares a workload group representing a collection of workloads that will be registered under reviews in namespace bookinfo. The set of labels will be associated with each workload instance during the bootstrap process, and the ports 3550 and 8080 will be associated with the workload group and use service account default. app.kubernetes.io/version is just an arbitrary example of a label.

apiVersion: networking.istio.io/v1
kind: WorkloadGroup
metadata:
  name: reviews
  namespace: bookinfo
spec:
  metadata:
    labels:
      app.kubernetes.io/name: reviews
      app.kubernetes.io/version: "1.3.4"
  template:
    ports:
      grpc: 3550
      http: 8080
    serviceAccount: default
  probe:
    initialDelaySeconds: 5
    timeoutSeconds: 3
    periodSeconds: 4
    successThreshold: 3
    failureThreshold: 3
    httpGet:
     path: /foo/bar
     host: 127.0.0.1
     port: 3100
     scheme: HTTPS
     httpHeaders:
     - name: Lit-Header
       value: Im-The-Best

WorkloadGroup

WorkloadGroup enables specifying the properties of a single workload for bootstrap and provides a template for WorkloadEntry, similar to how Deployment specifies properties of workloads via Pod templates. A WorkloadGroup can have more than one WorkloadEntry. WorkloadGroup has no relationship to resources which control service registry like ServiceEntry and as such doesn’t configure host name for these workloads.

Field	Description
`metadata` ObjectMeta	Metadata that will be used for all corresponding `WorkloadEntries`. User labels for a workload group should be set here in `metadata` rather than in `template`.
`template` WorkloadEntry Required	Template to be used for the generation of `WorkloadEntry` resources that belong to this `WorkloadGroup`. Please note that `address` and `labels` fields should not be set in the template, and an empty `serviceAccount` should default to `default`. The workload identities (mTLS certificates) will be bootstrapped using the specified service account’s token. Workload entries in this group will be in the same namespace as the workload group, and inherit the labels and annotations from the above `metadata` field.
`probe` ReadinessProbe	`ReadinessProbe` describes the configuration the user must provide for healthchecking on their workload. This configuration mirrors K8S in both syntax and logic for the most part.

ObjectMeta

ObjectMeta describes metadata that will be attached to a WorkloadEntry. It is a subset of the supported Kubernetes metadata.

Field	Description
`labels` map<string, string>	Labels to attach
`annotations` map<string, string>	Annotations to attach

ReadinessProbe

Field	Description
`initialDelaySeconds` int32	Number of seconds after the container has started before readiness probes are initiated.
`timeoutSeconds` int32	Number of seconds after which the probe times out. Defaults to 1 second. Minimum value is 1 second.
`periodSeconds` int32	How often (in seconds) to perform the probe. Default to 10 seconds. Minimum value is 1 second.
`successThreshold` int32	Minimum consecutive successes for the probe to be considered successful after having failed. Defaults to 1 second.
`failureThreshold` int32	Minimum consecutive failures for the probe to be considered failed after having succeeded. Defaults to 3 seconds.
`httpGet` HTTPHealthCheckConfig (oneof)	`httpGet` is performed to a given endpoint and the status/able to connect determines health.
`tcpSocket` TCPHealthCheckConfig (oneof)	Health is determined by if the proxy is able to connect.
`exec` ExecHealthCheckConfig (oneof)	Health is determined by how the command that is executed exited.
`grpc` GrpcHealthCheckConfig (oneof)	GRPC call is made and response/error is used to determine health.

HTTPHealthCheckConfig

Field	Description
`path` string	Path to access on the HTTP server.
`port` uint32 Required	Port on which the endpoint lives.
`host` string	Host name to connect to, defaults to the pod IP. You probably want to set “Host” in httpHeaders instead.
`scheme` string	HTTP or HTTPS, defaults to HTTP
`httpHeaders` HTTPHeader[]	Headers the proxy will pass on to make the request. Allows repeated headers.

GrpcHealthCheckConfig

Field	Description
`port` uint32	Port on which the endpoint lives.
`service` string	Service is the fully qualified name of the service to send the grpc health check request

HTTPHeader

Field	Description
`name` string	The header field name
`value` string	The header field value

TCPHealthCheckConfig

Field	Description
`host` string	Host to connect to, defaults to localhost
`port` uint32 Required	Port of host

ExecHealthCheckConfig

Field	Description
`command` string[] Required	Command to run. Exit status of 0 is treated as live/healthy and non-zero is unhealthy.