K8S Performance Optimization - K8S Node Parameter Tuning

preface

This is the fourth article in a series on K8S performance optimization: Best practices for Kubernetes node performance optimization parameters.

Series:

1. K8S Performance Optimization - OS sysctl Tuning
2. 《K8S Performance Optimization - K8S APIServer Tuning》
3. K8S Performance Optimization - CIDR Configuration for Large Clusters

Two parameters

Two parameters that control the maximum number of pods that can be scheduled for a K8S Node: podsPerCore and maxPods。

When both parameters are set, the smaller value limits the number of pods on the node. Exceeding these values can cause:

• CPU usage increases.
• Slow down pod scheduling.
• Depending on the amount of memory in the node, memory exhaustion can occur.
• Exhaust the IP address pool.
• Overuse of resources, resulting in poor user application performance.

⚠️ significant

In Kubernetes, a pod with a single container actually uses two containers. The second container is used to set up pause before the actual container starts. So, a system running 10 pods would actually run 20 containers.

podsPerCore Set the number of pods that a node can run based on the number of processor cores in the node. For example: on a node with 4 processor cores podsPerCore Set to 10 to allow the maximum number of pods allowed on that node to be 40.

kubeletConfig:
  podsPerCore: 10

will podsPerCore Set to 0 You can disable this restriction. The default is 0。podsPerCore Can’t exceed maxPods。

maxPods Set the number of pods that a node can run to a fixed value, regardless of the node’s properties.

kubeletConfig:
   maxPods: 250

EOF