Building and deploying scalable containerized applications with Kubernetes and EKS (Unknown) (z-library.sk, 1lib.sk, z-lib.sk)

【One-Line Pitch】 A hands-on field guide for engineers who already know containers and want to run Kubernetes on AWS properly—covering EKS architecture, cluster provisioning, storage, security, and cost-aware operations. Best suited to DevOps engineers, cloud architects, and platform administrators moving from tutorials to production. 【Book Arc】 - **Opening (~0%–10%)**: Traces container orchestration history from virtualization through Linux containers and Docker to Kubernetes' CNCF donation, then surveys adoption and enterprise maturity—establishing why Kubernetes became the default before any AWS specifics. - **Early (~10%–32%)**: Builds the Kubernetes mental model—controllers, labels, pods, services, scheduling, container runtimes and OCI, and storage primitives (volumes, PVs, PVCs)—then pivots to EKS itself: managed control plane, worker nodes, and the EC2-versus-Fargate choice. - **Middle (~32%–48%)**: Moves into practice: provisioning clusters via console, eksctl config files, and IaC tools (Terraform, CDK), plus RBAC and permission design, with worked examples and exercises. - **Late (~48%–onward)**: Covers day-two operations—cluster upgrades and upgrade insights, monitoring options (Container Insights, Kubecost), logging and troubleshooting—and applied patterns like ML model serving on GPU nodes. - **Ending**: Closes with optimization themes—performance, security, and cost-effectiveness—framed as the payoff for the earlier architecture and operations material. 【Key Takeaways】 - **EKS splits responsibility cleanly** (Early): AWS runs the control plane (API server, etcd, scheduling) across availability zones while you own worker nodes—this division is the lens for every later cost, security, and scaling decision. - **Worker node choice is a real trade-off** (Early): EC2 instances offer customization; Fargate offers serverless simplicity. The book frames the decision around cost, scalability, and operational overhead rather than a single "right" answer. - **Storage is designed around separation of concerns** (Early): Developers declare needs via PVCs; administrators provision PVs with reclaim policies and expansion—loose coupling that keeps storage portable across environments. - **Runtime flexibility comes from OCI** (Early): Docker, containerd, CRI-O, and specialized runtimes like Kata all execute the same workloads, reducing lock-in and enabling confidential-computing scenarios. - **Provisioning has a maturity ladder** (Middle): Console click-through for learning, eksctl config files for repeatability, and Terraform/CDK for fully automated, code-defined clusters—pick by team maturity, not preference. - **RBAC and least privilege are foundational, not optional** (Middle): Permissions for Kubernetes resources and API access are treated as a core design concern, reinforced later by tightly scoped IAM roles and DevSecOps integration. - **Upgrades need foresight** (Late): Upgrade insights surface deprecated APIs before version bumps, turning cluster maintenance from reactive breakage into planned transitions. - **Observability and cost are paired concerns** (Late): Container Insights for health, Kubecost for spend—monitoring is framed as both reliability and budget control. 【Reading Tips】 - Skim the history and CNCF chapters (~0%–10%) if you already know Kubernetes' origins; slow down at the EKS architecture and storage sections, where the book's real value sits. - Treat the eksctl/Terraform/CDK examples as templates—type them out and adapt rather than reading passively; the exercises at chapter ends are worth doing. - Deep-read the RBAC, upgrade, and monitoring chapters if you operate clusters; these are where production incidents originate. - Watch for the recurring cost/performance/security triad—it ties otherwise separate chapters together and is the book's implicit thesis. 【Coverage Limits】 The excerpts are fragmented (including OCR artifacts and table-of-contents fragments), so later chapters on monitoring, logging, and optimization are only partially represented; specific depth on networking internals and CI/CD pipelines is not verifiable from the sample.