與大家分享我在 COSCUP 2021 擔任 Cloud Native 議程講者的一些內容簡介與感想。雖然這次是線上議程，但大家在各平台上的積極發問與留言實在令人感到社群滿滿的活力。
這次我分享的主題是 “Empower Your Kubernetes with Service Mesh + Distributed Tracing”，主要介紹了 service mesh 與 distributed tracing 的基礎概念，並帶領大家以 kubernetes-native 的方式實際部署一個範例應用，將 distributed tracing 與 service mesh 無縫整合。
We will build a continuous delivery pipeline that deploys a React application to Amazon S3 and syncs it with Amazon CloudFront, a content delivery network (CDN) managed by AWS.
All source codes are available at https://github.com/minghsu0107/sync-react-s3-cloudfront.
Traefik is a powerful ingress controller with easy deployment and configuration. However, cloud providers like AWS and GCP also provide ingress implementations of their managed Kubernetes.
In this post, we will take GCP as example and walk through all needed knowledge of integrating Traefik deployment with ingress provided by GKE. This way, we could enjoy the benefits of feature-rich Traefik CRD as well as convenient infrastruture provisions provided by cloud ingress.
All source codes are available at https://github.com/minghsu0107/traefik-ingress-gke.
We are going to deploy a sample app container on Amazon ECS using Fargate, a serverless compute engine for containers. Fargate removes the need to provision and manage servers, lets us specify and pay for resources per application, and improves security through application isolation by design.
Amazon CloudFront is a fast content delivery network (CDN) service managed by AWS. It serves your contents across edge locations around the globe with high transfer speeds and low latency under secured connections.
In this post, we will set up an Amazon CloudFront distribution that serves private contents on your S3 bucket in order to speed up your content retrival while fully controlling user access permissions.
The following diagram shows a typical high-availibility Kubernetes cluster with embedded etcd:
However, HA architecture also brings about another problem - which K8s API server should the client connect to? How to detect a node failure while fowarding traffic in order to achieve high availibility?
在運行程式時，我們常希望善用系統的多核環境平行處理資料以提升計算的效率。然而 Python 的平行化似乎沒有那麼簡單。這篇文章會簡單介紹平行運算的概念、平行化 Python 時會遇到的問題、以及如何解決它們。
When developing HTTP APIs, we may have to process the same request-specific data throughout middlewares. Since it’s a quite common pattern, I decide to figure it out and share how I solve it.
這篇文章統整了我在 Golang Taipei #55 Meetup 分享的內容。
Event-driven architecture 在近幾年越來越受關注，它不僅幫助我們解耦服務組件、反轉依賴，更可提高系統的 throughput，大幅提升了擴展性。
這次主題會講解 Event-driven 的核心概念，簡介幾種常見的分佈式消息系統，並展示如何輕鬆用 Golang 實作 event-driven application，幫助大家能更快理解。
這個教學使用 DroneCI 與 ArgoCD 打造 cloud-native 的持續整合交付平台，讓我們在 push commit 或 merge PR 後即可自動跑完測試、打包 image 並部署到 K8s 叢集。必且藉由版控，我們也得以輕鬆 rollback 到之前的任一版本！
以下與大家分享一些這次 JCConf 2020 中我認為值得討論的主題。
整體來說，滿多議程在推廣 Kotlin 、serverless 與微服務的使用，而大家滿心期待的 Java virtual thread 也已如火如荼的開發中！
用 Hugo + Github Pages + Github Actions 快速架設部落格網站。
