User Issue: CPU pressure does not consume 100% of the pods allocated CPU

[sasidharsmit]

Describe the bug "CPU pressure" attack does not consume 100% of the CPU allocated to a Kubernetes pod.

To Reproduce Steps to reproduce the behavior: I injected a "CPU pressure" fault into a pod with 2 containers:

1. Application container (name: fault-injection-showcase, CPU requests: 16 cores, CPU limits: 16 cores)
2. Istio sidecar container (name: istio-proxy, CPU requests: 100 millicores, CPU limits: 2 cores)

Expected behavior The pod should consume 18 cores (100% of the CPU allocated to both the containers in the pod)

Actual behavior

1. The pod CPU usage fluctuates anywhere between 7 to 18 cores
2. Within the pod, the "istio-proxy" container was constantly at 100% CPU usage but the "fault-injection-showcase" container was only using ~60% all the time

Screenshots

Environment:

• Kubernetes version: v1.21.9
• Controller version: 6.0.0
• Cloud provider (or local): AWS
• Base OS for Kubernetes: Amazon Linux (5.4.181-99.354.amzn2.x86_64)