Tag: #open-source-internals
Writing on GPUs, LLMs, MLOps, Kubernetes — and mindset · 40 posts
etcd and Kubernetes: API Server Integration Analysis
Analyzes etcd and Kubernetes API Server integration. Covers how kube-apiserver uses etcd (storage backend, codec, transformer), key hierarchy (/registry/ prefix), API Server Watch Cache (cacher, watchCache, store), etcd
2026-03-20 · 3 min read #etcd#distributed-systems#open-source-internals[Prometheus] Service Discovery Mechanism: Complete Analysis
Complete analysis of Prometheus service discovery internals. Covers Discovery Manager architecture, Provider interface, kubernetessd watch mechanism, relabeling internals, filesd and HTTP SD, and the target lifecycle at
2026-03-20 · 9 min read #prometheus#observability#open-source-internals[Prometheus] Alerting Pipeline: From Rule Evaluation to Alertmanager Delivery
Deep dive into the Prometheus alerting pipeline internals. Covers Rule Manager evaluation mechanism, Alert state machine (inactive/pending/firing/resolved), for duration, Alertmanager internals (routing tree, inhibition,
2026-03-20 · 10 min read #prometheus#observability#open-source-internalsHelm Release Lifecycle: install, upgrade, rollback Internals
Analyze the complete Helm release lifecycle. Covers Secret-based release storage, install/upgrade/rollback flows, Hook system (pre-install, post-install, etc.), Three-way Strategic Merge Patch, atomic installs, and the -
2026-03-20 · 3 min read #helm#kubernetes#open-source-internalsHelm Architecture Internals: Design Philosophy and Structure of Helm 3
Deep dive into Helm 3 architecture internals. Covers the client-only architecture after Tiller removal, Secret/ConfigMap-based release storage, chart structure (Chart.yaml, values.yaml, templates/), Go template engine, a
2026-03-20 · 7 min read #helm#kubernetes#open-source-internals[containerd] Networking and Storage
Analyzes containerd networking and storage internals. Covers CNI integration, network namespaces, volume mounts, device access (GPU), SELinux/AppArmor at the source code level.
2026-03-20 · 19 min read #containerd#container-runtime#open-source-internals[containerd] Container Lifecycle Management
Analyzes containerd container lifecycle management. Covers container creation, OCI runtime spec, Task execution, shim lifecycle, checkpoint/restore, and runtime classes (runc, kata, gVisor, Wasm) at the source code level
2026-03-20 · 15 min read #containerd#container-runtime#open-source-internals[containerd] CRI Implementation: Kubernetes Runtime Integration
Analyzes containerd CRI implementation internals. Covers CRI gRPC service, Pod Sandbox, container spec translation, streaming (exec/attach/port-forward), RuntimeClass, and NRI (Node Resource Interface) at the source code
2026-03-20 · 18 min read #containerd#container-runtime#open-source-internalscontainerd Architecture Internals
Analyzes the architecture of containerd, the industry-standard container runtime. Covers gRPC API, plugin system (services, snapshotter, runtime), plugin registration/initialization/dependency resolution, CRI implementat
2026-03-20 · 3 min read #containerd#container-runtime#open-source-internalsCilium Network Policy Engine: L3/L4/L7 Filtering Implementation
Analysis of the Cilium network policy engine covering policy repository, Identity-based enforcement, L3/L4 eBPF enforcement, L7 Envoy proxy, FQDN policies, audit mode, and endpoint regeneration.
2026-03-20 · 9 min read #cilium#ebpf#networking#open-source-internalsCilium Hubble Observability Platform Internal Analysis
Internal analysis of the Cilium Hubble observability platform covering architecture, Relay, UI, L3/L4/L7 flows, Prometheus metrics, CLI, gRPC API, and performance tuning.
2026-03-20 · 9 min read #cilium#ebpf#networking#open-source-internalsCilium ClusterMesh: Multi-Cluster Networking Internal Implementation
Analysis of Cilium ClusterMesh internals covering etcd architecture, global services, cross-cluster policies, Identity synchronization, KVStoreMesh, and external workloads.
2026-03-20 · 7 min read #cilium#ebpf#networking#open-source-internalsCilium Architecture Internals: The Core of eBPF-Based Networking
Deep analysis of Cilium internals including the Cilium Agent, Operator, eBPF datapath, BPF maps, Identity allocation, and endpoint lifecycle.
2026-03-20 · 10 min read #cilium#ebpf#networking#open-source-internalsCilium eBPF Datapath Deep Dive: Packet Processing Pipeline
Deep analysis of the Cilium eBPF datapath covering packet flow, BPF program chaining, connection tracking, NAT engine, DSR mode, and socket-level load balancing.
2026-03-20 · 11 min read #cilium#ebpf#networking#open-source-internals[ArgoCD] Sync Engine Analysis: Everything About Synchronization Mechanisms
Analyzing the core mechanisms of the ArgoCD Sync engine. Covers Sync Phases, Resource Hooks, Wave ordering, Pruning, Diff engine, Health Checks, and Retry strategies in detail.
2026-03-20 · 9 min read #argocd#gitops#open-source-internals[ArgoCD] Architecture Internals: Source Code Level Deep Dive
A source-code-level analysis of ArgoCD core components: API Server, Repository Server, Application Controller, Redis, and Dex.
2026-03-20 · 10 min read #argocd#gitops#open-source-internals[ArgoCD] Security Model: Authentication, Authorization, and Secret Management
Deep analysis of the ArgoCD security model. Covers authentication (SSO, OIDC, LDAP), RBAC authorization, repository/cluster credentials, secret management, and network security.
2026-03-20 · 7 min read #argocd#gitops#open-source-internals[ArgoCD] ApplicationSet Controller Deep Dive
Deep dive into the ArgoCD ApplicationSet controller. Covers Generator types, template rendering, Progressive Sync, Go templating, and Reconciliation Loop in detail.
2026-03-20 · 7 min read #argocd#gitops#open-source-internals[containerd] Image Management: OCI Images and Snapshots
Analyzes the internals of containerd image management. Covers OCI image spec, Content Store, Snapshotter (overlayfs, native, devmapper), image pull flow, and garbage collection at the source code level.
2026-03-20 · 16 min read #containerd#container-runtime#open-source-internals[ArgoCD] High Availability and Scalability: Large-Scale Cluster Operations
Analyzing ArgoCD HA and scalability. Covers Controller sharding, Repo Server scaling, Redis HA, performance tuning, monitoring, and disaster recovery strategies.
2026-03-20 · 6 min read #argocd#gitops#open-source-internals