Monitoring APIRequestCount in OpenShift
2 min read | by Jordi Prats
Openshift provides an object that tracks the number of requests made to the Kubernetes API server. It provides insights into the load on the cluster, the performance of applications, and helps in capacity planning. By monitoring APIRequestCount, you can identify potential bottlenecks, detect unusual spikes in traffic, and optimize resource allocation.
$ kubectl get apirequestcounts NAME REMOVEDINRELEASE REQUESTSINCURRENTHOUR REQUESTSINLAST24H alertmanagerconfigs.v1alpha1.monitoring.coreos.com 6 1706 alertmanagers.v1.monitoring.coreos.com 20 2891 apiservices.v1.apiregistration.k8s.io 994 99521 (...) We can get the details for each of the available APIs. For example, we can use the following command to get stats about the calls issued to Pod. It will fetch the APIRequestCount data for Pods from the Kubernetes API server and display the results. The output will include information such as the API group, resource (Pods), version, namespace, and the number of requests made for each specific Pod:
$ kubectl get apirequestcounts pods.v1 NAME REMOVEDINRELEASE REQUESTSINCURRENTHOUR REQUESTSINLAST24H pods.v1 2161 137560 We can get an even more detailed response by retrieving the whole object using -o yaml:
$ kubectl get apirequestcounts pods.v1 -o yaml apiVersion: apiserver.openshift.io/v1 kind: APIRequestCount metadata: creationTimestamp: "2022-07-08T19:56:30Z" generation: 1 name: pods.v1 resourceVersion: "108979491" uid: 041cb88b-9d93-a411-dead-72a5bf79fa94 spec: numberOfUsersToReport: 10 status: currentHour: byNode: - byUser: - byVerb: - requestCount: 87 verb: get - requestCount: 21 verb: list - requestCount: 2 verb: watch requestCount: 110 userAgent: Go-http-client/2.0 username: system:serviceaccount:openshift-etcd-operator:etcd-operator - byVerb: - requestCount: 109 verb: watch requestCount: 109 userAgent: Prometheus/2.32.1 username: system:serviceaccount:openshift-monitoring:prometheus-k8s - byVerb: - requestCount: 34 verb: get requestCount: 34 userAgent: Mozilla/5.0 username: cluster-admin - byVerb: - requestCount: 26 verb: get - requestCount: 6 verb: update requestCount: 32 userAgent: multus/v0.0.0 username: system:serviceaccount:openshift-multus:multus - byVerb: - requestCount: 7 verb: create - requestCount: 14 verb: delete - requestCount: 6 verb: watch requestCount: 27 userAgent: catalog/v0.0.0 username: system:serviceaccount:openshift-operator-lifecycle-manager:olm-operator-serviceaccount By regularly monitoring APIRequestCount, setting up alerts, and analyzing the data over time, you can ensure the smooth operation of your applications and make informed decisions about resource allocation
Posted on 11/07/2023