Researches on Lê Tuấn Anh — Go Backend Architect & Microservices Engineerhttps://tanhdev.com/research/Recent content in Researches on Lê Tuấn Anh — Go Backend Architect & Microservices EngineerLê Tuấn Anh — Go Backend Architect & Microservices Engineerhttps://tanhdev.com/vesviet.pnghttps://tanhdev.com/vesviet.pngHugoenThu, 06 Aug 2026 00:00:00 +0000Building Custom Kubernetes Operators in Go with kubebuilder & Deep eBPF Kernel Observability using cilium/ebpfhttps://tanhdev.com/research/building-custom-kubernetes-operators-ebpf-golang-cilium/Thu, 06 Aug 2026 00:00:00 +0000https://tanhdev.com/research/building-custom-kubernetes-operators-ebpf-golang-cilium/An authoritative, production-grade research dossier on building custom Kubernetes v4 operators in Go integrated with deep eBPF kernel observability via cilium/ebpf, bpf2go, and zero-copy BPF ringbuffers.High-Throughput Local LLM Infrastructure: Architecting a Distributed Go API Gateway for vLLM & PagedAttention Clustershttps://tanhdev.com/research/high-throughput-local-llm-infrastructure-vllm-golang-gateway/Thu, 06 Aug 2026 00:00:00 +0000https://tanhdev.com/research/high-throughput-local-llm-infrastructure-vllm-golang-gateway/A comprehensive technical research dossier on architecting a high-throughput, low-latency distributed Go API Gateway for vLLM inference clusters. Features PagedAttention memory internals, Prefill-Decode disaggregation over RoCE v2/NVLink, prompt prefix context-affinity routing, empirical H100 vs L40S vs H200 benchmarks, 12-month SaaS vs self-hosted TCO analysis, complete Kubernetes manifests, and GPU VRAM sizing formulas.Modern Go 1.23/1.24 High-Performance Engineering: Custom Iterators (iter.Seq), Zero-Allocation Memory Pools, and Microsecond GC Tuninghttps://tanhdev.com/research/modern-golang-123-124-high-performance-zero-alloc-gc-tuning/Thu, 06 Aug 2026 00:00:00 +0000https://tanhdev.com/research/modern-golang-123-124-high-performance-zero-alloc-gc-tuning/Comprehensive deep-dive research dossier on Go 1.23 Range-Over-Func iterators, Go 1.24 string/struct interning with unique.Handle, escape analysis mechanics, multi-tiered sync.Pool memory buffers, and Kubernetes GOMEMLIMIT microsecond GC tuning.Production AI Observability: Building Zero-Overhead LLM Tracing & Cost Attribution with OpenTelemetry in Gohttps://tanhdev.com/research/production-ai-observability-opentelemetry-golang-llm-tracing/Thu, 06 Aug 2026 00:00:00 +0000https://tanhdev.com/research/production-ai-observability-opentelemetry-golang-llm-tracing/A complete engineering research dossier on building zero-overhead, OpenTelemetry-native LLM tracing, streaming TTFT/TPOT latency measurement, multi-agent W3C context propagation, and OTTL token cost attribution in Go.