Procedure

1. To ensure the stability of critical apps during peak hours, you need to consistently monitor their CPU and memory usage, identify and respond to peak values, and configure CPU and memory usage alerts for VMs.

2. The customer has CPU overcommitment set to unlimited for nodes. Therefore, you need to consistently monitor the CPU usage of the nodes to avoid persistent CPU preemption among VMs in the cluster. Otherwise, the performance of critical businesses will be affected. When the CPU usage of a node reaches 85% or higher and remains for 1 hour, critical alerts are triggered, and notifications are sent.