Wait on Resource Status for Applied Configuration

This feature is in the Alpha stage, see Istio Feature Status. Your feedback is welcome in the Istio User Experience discussion. Currently, this feature is tested only for single, low volume clusters with a single control plane revision.

Istio’s mesh configuration is declarative, which means that you define a configuration and Istio propagates the changes through the mesh over time. As a result, your command might attempt to use your service mesh before the relevant resources are ready.

In Istio 1.6 and later, you can use the kubectl wait command to have more control over the way that Istio applies configuration changes to the mesh. To make this possible, the kubectl wait command monitors the status field of the resource’s status, which Istio updates as it propagates configuration changes.

Before you begin

This feature is off by default. Enable the status field as part of Istio installation using the following command. You must also enable config_distribution_tracking.

$ istioctl install --set values.pilot.env.PILOT_ENABLE_STATUS=true --set values.pilot.env.PILOT_ENABLE_CONFIG_DISTRIBUTION_TRACKING=true --set values.global.istiod.enableAnalysis=true

Wait for resource readiness

You can apply a change and then wait for completion. For example, to wait for a virtual service, use the following commands:

ZipZip

$ kubectl apply -f @samples/httpbin/httpbin.yaml@
$ kubectl apply -f @samples/httpbin/httpbin-gateway.yaml@
$ kubectl wait --for=condition=Reconciled virtualservice/httpbin
virtualservice.networking.istio.io/httpbin condition met

This blocking command does not release until the virtual service has been distributed to all proxies in the mesh, or until the command times out.

When you use the kubectl wait command in a script, the return code will be 0 for success, or a non-zero value for time out.

For more information about usage and syntax, see the kubectl wait command.