The Metrics Server component uses an underlying HTTP/2 library. When network jitter occurs, this library may return the error http2: no cached connection was available. This error can cause several issues, such as the kubectl top node/pod command returning no output, the Horizontal Pod Autoscaler (HPA) failing to work, or namespace deletion being blocked.
Impact scope
You can resolve this issue by upgrading the component. This issue is fixed in versions v0.2.3-8a48069-aliyun and later. If you are using one of these versions, no upgrade is required.
Affected versions: v0.2.2-473c961-aliyun and v0.2.2-b4bf266-aliyun
Fixed versions: v0.2.3-8a48069-aliyun and later
Upgrade the component
Log on to the ACK console. In the left navigation pane, click Clusters.
On the Clusters page, click the name of your cluster. In the left navigation pane, click Add-ons.
On the Add-ons page, click the Log & Monitoring tab, locate the metrics-server component, click Upgrade, and then follow the prompts to complete the upgrade.
After the upgrade is successful, the component card displays Upgraded.