49 lines
2.1 KiB
Markdown
49 lines
2.1 KiB
Markdown
---
|
||
title: "Public Cloud Learning Sessions - EKS Optimization Part 1 of 3 - Compute Optimization with Karpenter"
|
||
type: source
|
||
tags: [AWS, EKS, Karpenter, Cost-Optimization, Compute-Optimization]
|
||
date: 2026-04-14
|
||
---
|
||
|
||
## Source File
|
||
- [[raw/Cloud & DevOps/Public-Cloud-Learning-Sessions/04_EKS/public-cloud-learning-sessions-eks-optimization-part-1-of-3-compute-optimization.md]]
|
||
|
||
## Summary
|
||
- 核心主题：EKS 计算优化，使用 Karpenter 实现自动扩缩容
|
||
- 问题域：Kubernetes 集群资源管理、成本优化
|
||
- 方法/机制：Karpenter 替代传统 Cluster Autoscaler，直接与 EC2 Fleet API 通信，基于成本和利用率进行工作负载放置和整合
|
||
- 结论/价值：简化数据平面管理，降低节点管理复杂度，实现更优的成本效率
|
||
|
||
## Key Claims
|
||
- Karpenter 与 Kubernetes workload scheduling constructs 原生集成
|
||
- Karpenter 直接与 EC2 Fleet API 通信，降低延迟
|
||
- Karpenter 提供工作负载放置和节点整合的原生体验
|
||
- Karpenter 原生支持 Spot 中断处理，使用 EventBridge 和 SQS
|
||
|
||
## Key Quotes
|
||
> "Carpenter not only does the auto-scaling bit, but it also removes the pain points of working with node groups."
|
||
|
||
> "Carpenter has native integration with Kubernetes and it complements the native Kubernetes spot pod scheduling constraints that is available for your workloads."
|
||
|
||
## Key Concepts
|
||
- [[Karpenter]]：开源 Kubernetes compute management tool，替代 Cluster Autoscaler
|
||
- [[Node-Pools]]：定义调度约束和容量限制
|
||
- [[Node-Classes]]：定义实例配置细节（子网、节点角色、AMI）
|
||
- [[Spot-Interruption]]：Spot 实例中断处理
|
||
- [[Consolidation-Policies]]：整合策略，控制成本优化行为
|
||
|
||
## Key Entities
|
||
- [[AWS]]：云服务提供商
|
||
- [[EKS]]：Amazon Elastic Kubernetes Service
|
||
- [[EC2-Fleet-API]]：AWS EC2 灵活计算实例 API
|
||
- [[EventBridge]]：AWS 事件总线服务
|
||
- [[SQS]]：AWS 简单队列服务
|
||
|
||
## Connections
|
||
- [[Karpenter]] ← extends ← [[Cluster-Autoscaler]]
|
||
- [[Karpenter]] ← integrates_with ← [[EC2-Fleet-API]]
|
||
- [[Karpenter]] ← uses ← [[EventBridge]]
|
||
- [[EKS]] ← manages ← [[Karpenter]]
|
||
|
||
## Contradictions
|
||
- （暂无） |