kubernetes Index

📙 kubernetes Index

📚 32 Posts
🕒 Last Updated: Mon Aug 17 2026

This folder contains kubernetes-related posts.

#	Blog Link	Date	Excerpt	Tags
1	kubernetes Index	Mon Aug 17 2026	📙 Index of kubernetes posts
2	Kubernetes: Control Loops, Scheduling, and GPUs	Fri Feb 20 2026	A working engineer's map of Kubernetes — the reconciliation model underneath the objects, the scheduling and networking internals that bite at scale, and how GPU workloads actually get placed and run.	Kubernetes DevOps Cloud Containers Orchestration GPU Scheduling MLOps
3	Image Internals: From OCI Layers to a Running Container	Fri Jul 24 2026	The complete journey of a container image — OCI manifest, content-addressable layers, registry pull flow, containerd snapshots, OverlayFS rootfs assembly, pod sandbox creation, and how a container process finally starts inside a pod.	Kubernetes Containers OCI Docker containerd OverlayFS Container Runtime Image Registry DevOps
4	Container Internals: What a Container Really Is	Wed Jul 22 2026	What a container actually is from the OS and Kubernetes perspectives — Linux namespaces, cgroups, OverlayFS image layers, the OCI spec, the container runtime stack from CRI to runc, and container security primitives.	Kubernetes Containers Linux Docker OCI OverlayFS Container Runtime Security DevOps
5	Kubernetes Pod Internals: What a Pod Really Is	Wed Jul 22 2026	What a Kubernetes pod actually is from the OS and Kubernetes perspectives — Linux namespaces, cgroups, the pause container, shared networking, pod lifecycle, init containers, and what happens between kubectl apply and your process running.	Kubernetes Pod Linux Namespaces cgroups Container Runtime DevOps Cloud
6	Kubernetes API Server Internals	Tue Jul 07 2026	Deep dive into the Kubernetes API Server — authentication, authorization, RBAC, admission controllers, schema validation, the watch cache, optimistic concurrency, and API Priority & Fairness explained with diagrams.	Kubernetes API Server Control Plane Security RBAC DevOps Cloud
7	etcd Architecture Explained	Tue Jul 07 2026	etcd internals for Kubernetes engineers — Raft consensus, leader election, MVCC and resourceVersion, snapshots, log compaction, the watch API, quorum loss behavior, and etcdctl operations for backup and defrag.	Kubernetes etcd Raft Control Plane Distributed Systems Storage DevOps
8	Kubernetes Scheduler Internals	Tue Jul 07 2026	Inside the Kubernetes Scheduler — scheduling queue, filtering, scoring, preemption, binding, topology spread constraints, the scheduler plugin framework, and where Kueue fits for batch and AI workloads.	Kubernetes Scheduler Pod Scheduling Control Plane Kueue DevOps Cloud
9	Kubelet Internals: The Node Agent That Runs Everything	Wed Jul 29 2026	What the kubelet actually does on every node — the sync loop and pod sources, the Container Runtime Interface, the Pod Lifecycle Event Generator, node heartbeats and leases, static pods, cgroup enforcement, the eviction manager, probe execution, and the Device Manager that hands out GPUs.	Kubernetes Kubelet Node Agent CRI Control Plane DevOps Cloud
10	Kubernetes Informers & Controllers Explained	Tue Jul 07 2026	How Kubernetes controllers and informers work — reconciliation loops, shared informers, local caches, work queues with exponential backoff, owner references, generation tracking, finalizers, and the operator pattern.	Kubernetes Controllers Informers Reconciliation Control Plane DevOps Cloud
11	Kubernetes Networking: Pods, Services, Ingress, and CNI	Tue Jul 07 2026	How Kubernetes networking works from the ground up — the flat Pod network model, CNI plugins, kube-proxy and iptables, Services (ClusterIP, NodePort, LoadBalancer, Headless), CoreDNS service discovery, Ingress, NetworkPolicies, and why overlay networks are replaced by InfiniBand for GPU training.	Kubernetes Networking CNI Services Ingress CoreDNS NetworkPolicy DevOps Cloud KCNA
12	Kubernetes Storage: PV, PVC, StorageClass, and CSI	Tue Jul 07 2026	How Kubernetes persistent storage works — Volumes vs PersistentVolumes, PersistentVolumeClaims, StorageClass dynamic provisioning, access modes, reclaim policies, the CSI driver model, StatefulSet stable storage, and storage requirements for GPU training checkpoints.	Kubernetes Storage PersistentVolume StorageClass CSI StatefulSet DevOps Cloud KCNA
13	Helm: Kubernetes Package Manager	Tue Jul 07 2026	Helm from the ground up — charts, releases, repositories, values, and Go templates. Essential commands, values overrides, chart structure, Helm hooks, Helm vs Kustomize, and real examples using GPU Operator, NIM, and Kueue.	Kubernetes Helm DevOps Cloud Package Manager GitOps GPU Operator KCNA
14	Cloud Native Observability: Prometheus, Grafana, OpenTelemetry, and Tracing	Tue Jul 07 2026	The three pillars of observability on Kubernetes — metrics with Prometheus and PromQL, visualization with Grafana, structured logging with Loki and Fluent Bit, distributed tracing with OpenTelemetry and Tempo, the OTel Collector pipeline, and GPU-specific observability with DCGM on DGX clusters.	Kubernetes Observability Prometheus Grafana OpenTelemetry Tracing Loki Logging DCGM DevOps Cloud KCNA
15	Kubernetes Resource Allocation: Requests, Limits, QoS, and Quotas	Wed Jul 22 2026	How Kubernetes allocates CPU and memory — requests vs limits, QoS classes, ResourceQuota, LimitRange, node allocatable capacity, GPU resources, and best practices for production workloads.	Kubernetes Resource Management QoS ResourceQuota LimitRange GPU DevOps Cloud
16	GPU Scheduling in Kubernetes: Device Plugins, GPU Operator & MIG	Tue Jul 07 2026	How Kubernetes schedules GPUs — Device Plugin gRPC protocol, nvidia-container-toolkit, GPU Operator ClusterPolicy, MIG profiles on H100, time-slicing, DCGM monitoring, GPU health tainting, and GPU sharing strategies for AI workloads.	Kubernetes GPU NVIDIA MIG GPU Operator AI MLOps DevOps Cloud
17	NVIDIA Network Operator: InfiniBand, SR-IOV, RDMA, and Multus	Tue Jul 07 2026	Why a DGX cluster trains 10× faster than a regular GPU cluster — the full network stack explained: RDMA, GPUDirect RDMA, InfiniBand, SR-IOV, Multus CNI, and how the NVIDIA Network Operator automates all of it on Kubernetes.	Kubernetes InfiniBand RDMA SR-IOV Multus NVIDIA Network Operator DGX Distributed Training MLOps DevOps
18	Dynamic Resource Allocation: The Future of GPU Scheduling in Kubernetes	Tue Jul 07 2026	How Kubernetes DRA replaces Device Plugins for GPU scheduling — ResourceClaim, DeviceClass, ResourceSlice, CEL selectors, topology-aware allocation, the NVIDIA GPU DRA driver, and how DRA and Device Plugins coexist during migration.	Kubernetes DRA GPU NVIDIA Scheduling DGX AI MLOps DevOps
19	Kubernetes Performance at Scale	Tue Jul 07 2026	Kubernetes at hyperscale — official SLIs and SLOs, pod startup latency breakdown, watch storms, LIST scalability, API Priority & Fairness, etcd bottlenecks, scheduler throughput, horizontal API Server scaling, and benchmarking with ClusterLoader2 and KWOK.	Kubernetes Performance Scalability APF etcd Control Plane KWOK ClusterLoader2 DevOps
20	Optimizing AI Inference at Scale: The Full Stack	Tue Jul 21 2026	There is no single technique to keep GPUs busy. A layer-by-layer map of AI inference optimization — from quantization and inference engines through KV cache, continuous batching, GPU sharing, scheduling, and cache-aware routing up to parallelism, autoscaling, and the networking underneath.	AI Infrastructure GPU Inference LLM Kubernetes MLOps vLLM Quantization Scheduling Autoscaling
21	Kueue: Kubernetes-Native Job Queuing and Quota Management	Tue Jul 07 2026	Deep dive into Kueue — the CNCF project that adds job queuing, resource quotas, gang scheduling, preemption, and fair sharing to Kubernetes. Covers ResourceFlavors, ClusterQueues, LocalQueues, Cohorts, and integration with PyTorchJob and batch workloads.	Kubernetes Kueue Job Scheduling GPU MLOps Batch DGX AI DevOps
22	Multi-Node Distributed Training on Kubernetes	Tue Jul 07 2026	How Kubernetes orchestrates distributed AI training across multiple DGX nodes — Kubeflow Training Operator, PyTorchJob, gang scheduling, NCCL, AllReduce, intra-node NVLink vs inter-node InfiniBand, and fault-tolerant checkpointing.	Kubernetes Distributed Training DGX PyTorch Kubeflow NCCL InfiniBand GPU AI MLOps
23	Kubernetes Topology Manager: NUMA-Aware GPU Scheduling	Tue Jul 07 2026	How the kubelet Topology Manager co-locates GPUs, CPUs, memory, and NICs on the same NUMA node — the difference between 1 μs and 100 ns RDMA latency on DGX. Covers NUMA basics, CPU Manager, Memory Manager, hint collection, and the four Topology Manager policies.	Kubernetes Topology Manager NUMA GPU DGX Performance CPU Manager InfiniBand RDMA MLOps DevOps
24	NVIDIA NIM: Optimized Inference Microservices on Kubernetes	Tue Jul 07 2026	What NVIDIA NIM is and how it works — NIM profiles, NGC model cache, the NIM Operator, NIMService CRD, OpenAI-compatible API, GPU-aware deployment on Kubernetes, and how NIM compares to raw Triton and vLLM for production inference on DGX Cloud.	Kubernetes NIM NVIDIA Inference LLM DGX AI MLOps TensorRT-LLM Triton
25	GPU Autoscaling on Kubernetes: KEDA, HPA, and Cluster Autoscaler	Tue Jul 07 2026	How to autoscale GPU workloads on Kubernetes — DCGM metrics pipeline to HPA, KEDA ScaledObjects with Prometheus triggers, Cluster Autoscaler for GPU node groups, scale-down protection for training jobs, and KEDA + Kueue integration for queue-depth-driven scaling.	Kubernetes Autoscaling GPU KEDA HPA Cluster Autoscaler DCGM NVIDIA DGX AI MLOps DevOps
26	Fine-Tuning LLMs: LoRA, QLoRA, PEFT, and NeMo on Kubernetes	Tue Jul 21 2026	Why fine-tuning exists, the memory math that makes full fine-tuning prohibitive, LoRA's low-rank decomposition trick, QLoRA on quantized base models, instruction tuning vs RLHF vs DPO, and how to run fine-tuning jobs on a DGX Kubernetes cluster with NeMo and PyTorchJob.	Kubernetes LoRA PEFT Fine-Tuning NeMo LLM NVIDIA DGX AI MLOps
27	Flash Attention: Fast, Memory-Efficient Attention for LLMs	Tue Jul 21 2026	How standard self-attention creates an O(N²) memory bottleneck, the IO-aware tiling algorithm that Flash Attention uses to stay in SRAM, Flash Attention 2 and 3 improvements, Grouped Query Attention and its KV cache impact, PagedAttention, and how these optimizations flow through TensorRT-LLM and NIM on H100.	Attention LLM NVIDIA CUDA TensorRT FlashAttention Inference AI Performance
28	Kubernetes and Cloud Native Certification Path	Tue Feb 24 2026	Foundational concepts of Kubernetes and the cloud native ecosystem, covering container orchestration, architecture, observability, and core Kubernetes components.	Kubernetes Cloud Native KCNA CNCF Containers DevOps Certification
29	KCNA Mock Exam — Set 1	Wed Jul 29 2026	A 60-question practice exam for the Kubernetes and Cloud Native Associate (KCNA) certification, weighted to the official exam domains — Kubernetes Fundamentals, Container Orchestration, Cloud Native Application Delivery, and Cloud Native Architecture.	Kubernetes KCNA CNCF Certification Mock Exam Cloud Native DevOps
30	KCNA Mock Exam — Set 2	Wed Jul 29 2026	A second 60-question practice exam for the Kubernetes and Cloud Native Associate (KCNA) certification, covering the same official exam domains with a fresh question set for additional practice.	Kubernetes KCNA CNCF Certification Mock Exam Cloud Native DevOps
31	KCSA Mock Exam — Set 1	Wed Jul 29 2026	A 60-question practice exam for the Kubernetes and Cloud Native Security Associate (KCSA) certification, covering the 4Cs security model, cluster component security, Kubernetes threat modeling, platform security, and compliance frameworks.	Kubernetes KCSA CNCF Security Certification Mock Exam Cloud Native DevOps
32	KCSA Mock Exam — Set 2	Wed Jul 29 2026	A second 60-question practice exam for the Kubernetes and Cloud Native Security Associate (KCSA) certification, covering the same official security domains with a fresh question set for additional practice.	Kubernetes KCSA CNCF Security Certification Mock Exam Cloud Native DevOps