OpenShift Serverless release notes

In previous versions of OpenShift Serverless, the KnativeServing custom resource shows a status of Ready, even if Kourier does not deploy. This bug is fixed in OpenShift Serverless 1.7.2.

OpenShift Serverless now uses Knative Serving 0.13.3.

OpenShift Serverless now uses Knative Serving Operator 0.13.3.

OpenShift Serverless now uses Knative kn CLI 0.13.2.

OpenShift Serverless uses Knative eventing 0.13.0.

OpenShift Serverless now uses Knative eventing Operator 0.13.3.

In OpenShift Serverless 1.7.0, routes were reconciled continuously when this was not required. This bug is fixed in OpenShift Serverless 1.7.1.

OpenShift Serverless 1.7.0 is now Generally Available (GA) on OpenShift Container Platform 4.3 and newer versions. In previous versions, OpenShift Serverless was a Technology Preview.

OpenShift Serverless now uses Knative Serving 0.13.2.

OpenShift Serverless now uses Knative Serving Operator 0.13.2.

OpenShift Serverless now uses Knative kn CLI 0.13.2.

Knative kn CLI downloads now support disconnected, or restricted network installations.

Knative kn CLI libraries are now signed by Red Hat.

Knative eventing is now available as a Technology Preview with OpenShift Serverless. OpenShift Serverless uses Knative eventing 0.13.2.

High availability (HA) is now enabled by default for the autoscaler-hpa, controller, activator , kourier-control, and kourier-gateway controllers.

OpenShift Serverless now supports the trustedCA setting in OpenShift Container Platform’s cluster-wide proxy, and is now fully compatible with OpenShift Container Platform’s proxy settings.

OpenShift Serverless now supports HTTPS by using the wildcard certificate that is registered for OpenShift Container Platform routes. For more information on HTTP and HTTPS on Knative Serving, see the documentation on Verifying your serverless application deployment.

In previous versions, requesting KnativeServing CRs without specifying an API group, for example, by using the command oc get knativeserving -n knative-serving, occasionally caused errors. This issue is fixed in OpenShift Serverless 1.7.0.

In previous versions, the Knative Serving controller was not notified when a new service CA certificate was generated due to service CA certificate rotation. New revisions created after a service CA certificate rotation were failing with the error:

Revision "foo-1" failed with message: Unable to fetch image "image-registry.openshift-image-registry.svc:5000/eap/eap-app": failed to resolve image to digest: failed to fetch image information: Get https://image-registry.openshift-image-registry.svc:5000/v2/: x509: certificate signed by unknown authority.

When upgrading from OpenShift Serverless 1.6.0 to 1.7.0, support for HTTPS requires a change to the format of routes. Knative services created on OpenShift Serverless 1.6.0 are no longer reachable at the old format URLs. You must retrieve the new URL for each service after upgrading OpenShift Serverless. For more information, see the documentation on Upgrading OpenShift Serverless.

If you are using Knative eventing on an Azure cluster, it is possible that the imc-dispatcher pod may not start. This is due to the pod’s default resources settings. As a work-around, you can remove the resources settings.

If you have 1000 Knative services on a cluster, and then perform a reinstall or upgrade of Knative Serving, there is a delay when you create the first new service after the KnativeServing CR becomes Ready.

OpenShift Serverless 1.6.0 is available on OpenShift Container Platform 4.3 and newer versions.

OpenShift Serverless now uses Knative Serving 0.13.1.

OpenShift Serverless now uses Knative kn CLI 0.13.1.

OpenShift Serverless now uses Knative Serving Operator 0.13.1.

The serving.knative.dev API group has now been fully deprecated and is replaced by the operator.knative.dev API group.

$ oc delete crd knativeservings.serving.knative.dev

The Subscription Update Channel in the web console for new OpenShift Serverless releases was updated from techpreview to preview-4.3.

OpenShift Serverless now supports the use of HTTP_PROXY.

OpenShift Serverless now supports HTTPS_PROXY cluster-proxy settings.

The KnativeServing CRD is now hidden from the Developer Catalog by default so that only users with cluster administrator permissions can view it.

Parts of the KnativeServing control plane and data plane are now deployed as highly available (HA) by default.

Kourier is now actively watched and reconciles changes automatically.

OpenShift Serverless now supports use on OpenShift Container Platform nightly builds.

In previous versions, the oc explain command did not work correctly. The structural schema of the KnativeServing CRD was updated in OpenShift Serverless 1.6.0 so that the oc explain command now works correctly.

In previous versions, it was possible to create more than one KnativeServing custom resource (CR). Multiple KnativeServing CRs are now prevented synchronously in OpenShift Serverless 1.6.0. Attempting to create more than one KnativeServing CR now results in an error.

In previous versions, OpenShift Serverless was not compatible with OpenShift Container Platform deployments on GCP. This issue was fixed in OpenShift Serverless 1.6.0.

In previous releases, the Knative Serving webhook crashed with an out of memory error if the cluster had more than 170 namespaces. This issue was fixed in OpenShift Serverless 1.6.0.

In previous releases, OpenShift Serverless did not automatically fix an OpenShift Container Platform route that it created if the route was changed by another component. This issue was fixed in OpenShift Serverless 1.6.0.

In previous versions, deleting a KnativeServing CR occasionally caused the system to hang. This issue was fixed in OpenShift Serverless 1.6.0.

Due to the ingress migration from Service Mesh to Kourier that occured in OpenShift Serverless 1.5.0, orphaned VirtualServices sometimes remained on the system. In OpenShift Serverless 1.6.0, orphaned VirtualServices are automatically removed.

In OpenShift Serverless 1.6.0, if a cluster administrator uninstalls OpenShift Serverless by following the uninstall procedure provided in the documentation, the Serverless dropdown is still be visible in the Administrator perspective of the OpenShift Container Platform web console, and the Knative Service resource is still be visible in the Developer perspective of the OpenShift Container Platform web console. Although you can create Knative services by using this option, these Knative services do not work.

$ oc get crd -oname | grep -e '(serving|internal).knative.dev' | xargs oc delete

OpenShift Serverless 1.5.0 is available on OpenShift Container Platform 4.3 and newer versions.

OpenShift Serverless has been updated to use Knative Serving 0.12.1.

OpenShift Serverless has been updated to use Knative kn CLI 0.12.0.

OpenShift Serverless has been updated to use Knative Serving Operator 0.12.1.

OpenShift Serverless ingress implementation has been updated to use Kourier in place of Service Mesh. No user intervention is necessary, as this change is automatic when the OpenShift Serverless Operator is upgraded to 1.5.0.

In previous releases, OpenShift Container Platform scale from zero latency caused a delay of approximately 10 seconds when creating pods. This issue has been fixed in the OpenShift Container Platform 4.3.5 bug fix update.

Deleting the KnativeServing.operator.knative.dev custom resource definition (CRD) from the knative-serving namespace can cause the deletion process to hang. This is due to a race condition between deletion of the CRD and the knative-openshift-ingress ingress removing finalizers.