TKE will be upgraded to provide more stable basic monitoring services and improve the availability of monitoring data, alarming, and HPA scaling. The upgrade falls into three phases, that is, deploying the new version of the monitoring service add-on, switching the add-on version, and deactivating the earlier version, without affecting the services.
tke-monitor-agent monitoring data collection add-on will be installed under the
kube-system namespace of your cluster. The agent will use less than 70 MB memory and 0.01 CPU cores on each node. For add-on details, see Description of tke-monitor-agent.
This phase will be performed one week after the new version of the add-on is deployed to ensure the data source stability.
metrics-servicewill be created under the
kube-systemnamespace and point to the
metrics-servermaintained on the container side.
hpa-metrics-servicein the cluster will point to the new data source. If
apiservice v1beta1.metrics.k8s.iouses the default
kube-system/hpa-metrics-servicedata source provided by the container, and the Kubernetes cluster is on v1.8 or later, it will switch to the more stable data source
apiservice v1alpha1.monitor.tencent.iowill be added to support TKE virtual nodes in reporting monitoring data and querying the monitoring data of Pods on each node.
This phase will be performed one week after the add-on version is switched.