Skip to content

kubeapi-priority.md

Latest commit

 

History

History
84 lines (53 loc) · 3.65 KB

kubeapi-priority.md

File metadata and controls

84 lines (53 loc) · 3.65 KB
copyright
years
2014, 2022
lastupdated 2022-01-11
keywords openshift
subcollection openshift

{{site.data.keyword.attribute-definition-list}}

Setting Kubernetes API priority and fairness

{: #kubeapi-priority}

Your {{site.data.keyword.openshiftlong}} clusters have default settings in place to process simultaneous requests to the API server and prevent traffic overload. You can configure your own flow schema and priority levels for requests that are made to the API server of your clusters. For more information, see API priority and fairness{: external} in the Kubernetes documentation. {: shortdesc}

For example, you might have a user or namespace that runs your critical apps in prod. You can create a flow schema and priority so that your critical apps have a higher priority for the API server to fulfill their requests than other apps in the cluster.

The Kubernetes API priority and feature gate is enabled in clusters that run {{site.data.keyword.openshiftshort}} version 4.6 or later. {: note}

Reviewing default flow schema and priority levels

{: #kubeapi-default-priority}

{{site.data.keyword.openshiftlong_notm}} sets certain default flow schema and priority levels in addition to the default settings from Kubernetes. {: shortdesc}

Flow schema Resources that requests come from Priority level
apiserver-health Kubernetes API server health resources Custom priority level for these resources.
calico-system-service-accounts Resources in the calico-system namespace that use a service account in the namespace Same priority as kube-system namespace service accounts
ibm-system-service-accounts Resources in the ibm-system namespace that use a service account in the namespace Same priority as kube-system namespace service accounts
tigera-operator-service-accounts Resources in the tigera-operator namespace that use a service account in the namespace Same priority as kube-system namespace service accounts
{: caption="Default flow schema and priority levels" caption-side="top"}
{: summary="The rows are read from left to right. The flow schema is listed in the first column, the resources that the requests come from are listed in the second column, and the priority level is listed in the third column."}

You can create your own flow schema and priorities, but don't modify the default settings. Unexpected results might occur in your cluster when you modify API request priorities. {: important}

Follow the steps to review the flow schemas and priority levels set by {{site.data.keyword.openshiftlong_notm}}.

  1. List all flow schemas in your cluster, including those set by {{site.data.keyword.openshiftlong_notm}}, and their corresponding priority levels .

    oc get flowschemas

    {: pre}

  2. Review the details of a particular flow schema including which resources can make prioritized API requests, what type of API requests can be made, and what objects the requests can modify.

    oc describe flowschema <flow-schema-name>

    {: pre}

Viewing {{site.data.keyword.openshiftlong_notm}} created priority level configurations

{: #kube-api-prioritylevelconfig}

{{site.data.keyword.openshiftlong_notm}} sets a custom priority level configuration for the apiserver-health resource. {: shortdesc}

Use the following commands to view details about the configuration.

oc get prioritylevelconfiguration apiserver-health

{: pre}

oc describe prioritylevelconfiguration apiserver-health

{: pre}