This guide walks cluster administrators through installing the OpenShift Serverless Operator to an OpenShift Container Platform cluster.

OpenShift Serverless is supported for installation in a restricted network environment. For more information, see Using Operator Lifecycle Manager on restricted networks.

Cluster sizing requirements

To run OpenShift Serverless, the OpenShift Container Platform cluster must be sized correctly. The minimum requirement to use OpenShift Serverless is a cluster with 10 CPUs and 40GB memory.

The total size requirements to run OpenShift Serverless are dependent on the applications deployed. By default, each pod requests ~400m of CPU, so the minimum requirements are based on this value.

In the size requirement provided, an application can scale up to 10 replicas. Lowering the actual CPU request of applications can increase the number of possible replicas.

You can use the MachineSet API to manually scale your cluster up to the desired size. The minimum requirements usually mean that you must scale up one of the default MachineSets by two additional machines.

For more information on using the MachineSet API, see the documentation on Creating MachineSets.

For more information on scaling a MachineSet manually, see the documentation on manually scaling MachineSets.

The requirements provided relate only to the pool of worker machines of the OpenShift Container Platform cluster. Master nodes are not used for general scheduling and are omitted from the requirements.

The following limitations apply to all OpenShift Serverless deployments:

  • Maximum number of Knative services: 1000

  • Maximum number of Knative revisions: 1000

Additional requirements for advanced use-cases

For more advanced use-cases such as logging or metering on OpenShift Container Platform, you must deploy more resources. Recommended requirements for such use-cases are 24 CPUs and 96GB of memory.

If you have high availability (HA) enabled on your cluster, this requires between 0.5 - 1.5 cores and between 200MB - 2GB of memory for each replica of the Knative Serving control plane. HA is enabled for some Knative Serving components by default. You can disable HA by following the documentation on Configuring High Availability components.

Before upgrading to the latest Serverless release, you must remove the community Knative Eventing operator if you have previously installed it. Having the Knative Eventing operator installed will prevent you from being able to install the latest Technology Preview version of Knative Eventing that is included with OpenShift Serverless 1.7.0.

Installing the OpenShift Serverless Operator

This procedure describes how to install and subscribe to the OpenShift Serverless Operator from the OperatorHub using the OpenShift Container Platform web console.

Procedure
  1. In the OpenShift Container Platform web console, navigate to the CatalogOperatorHub page.

  2. Scroll, or type they keyword Serverless into the Filter by keyword box to find the OpenShift Serverless Operator.

    OpenShift Serverless Operator in the OpenShift Container Platform web console
  3. Review the information about the Operator and click Install.

    OpenShift Serverless Operator information
  4. On the Create Operator Subscription page: