Azure AKSKubernetes

Node NotReady from Memory/CPU Saturation

critical
Resource ContentionUpdated Jun 27, 2025

AKS nodes enter NotReady state when memory or CPU saturation causes kubelet and containerd to become unresponsive. This occurs when resource limits are exceeded or PSI metrics indicate sustained pressure.

Sources
Proactive monitoring best practices for Azure Kubernetes Service (AKS) - Azure Kubernetes Servicelearn.microsoft.com
Identify high CPU utilization in AKS clusters - Azurelearn.microsoft.com
Troubleshoot CPU Pressure in AKS Clusters Using PSI Metrics - Azurelearn.microsoft.com
Troubleshoot memory saturation in AKS clusters - Azurelearn.microsoft.com
Technologies:
Azure AKSSymptoms of this issue are visible in Azure AKS metrics and logs
KubernetesThe root cause of this issue originates in Kubernetes
How to detect:

Detect when node status becomes NotReady for >20 minutes AND node memory working set percentage >90% OR node CPU usage percentage >90% OR PSI cpu waiting seconds increasing.

Recommended action:

Identify high-consuming pods, review resource requests/limits, enable Horizontal Pod Autoscaler, or scale node pool to higher SKU VMs. For CPU pressure specifically, consider removing CPU limits and relying on CPU shares.