[{"data":1,"prerenderedAt":829},["ShallowReactive",2],{"tag-labels":3,"site-header":4,"article-aplikasimu-lambat-dan-kamu-udah-kepikiran-ganti-bahasa-tunggu-dulu-id":72},{},{"featuredPost":5,"navigations":29},{"id":6,"title":7,"title_en":8,"slug":9,"excerpt":10,"excerpt_en":11,"has_translation":12,"published_at":13,"reading_time":14,"hidden_from_listing":15,"tags":16,"thumbnail_image":21,"og_image":22,"series":23,"authors":28},110,"Ketika Pilih Framework Cuma Biar Kelihatan Beda","When You Pick a Framework Just to Look Different","ketika-pilih-framework-cuma-biar-kelihatan-beda","Terakhir kali kamu milih framework atau library baru, itu karena masalahmu butuh alat itu, atau karena alatnya kedengaran keren dan belum banyak yang pakai? Pertanyaan ini nggak enak dijawab. Soalnya keputusan teknologi sering diambil dengan alasan yang nggak pernah ditulis di dokumen teknis mana pun.","The last time you chose a new framework or library, was it because your problem needed that tool, or because it sounded cool and few people around you were using it? That question is uncomfortable. Technology decisions are often made for reasons that never appear in any technical document.",true,"2026-10-06T15:25:21.000Z",5,false,[17,18,19,20],"karier","framework","opini","pengalaman",null,"\u002Fog\u002Fketika-pilih-framework-cuma-biar-kelihatan-beda.png",{"id":24,"title":25,"title_en":26,"slug":27},9,"Artikel","Articles","all",[],[30,46,67],{"collection":31,"item":32},"series",{"__typename":31,"title":25,"title_en":26,"slug":27,"children_series":33},[34,38,42],{"title":35,"title_en":36,"slug":37},"Kecerdasan Artifisial","Artificial Intelligence","kecerdasan-artifisial",{"title":39,"title_en":40,"slug":41},"Konsep Dasar Pemrograman","Programming Fundamentals","konsep-dasar-pemrograman",{"title":43,"title_en":44,"slug":45},"Algoritma dan Struktur Data","Algorithms and Data Structures","algoritma-struktur-data",{"collection":31,"item":47},{"__typename":31,"title":48,"title_en":48,"slug":49,"children_series":50},"Tutorial","tutorial",[51,55,59,63],{"title":52,"title_en":53,"slug":54},"Belajar Python","Learning Python","belajar-python",{"title":56,"title_en":57,"slug":58},"Belajar Golang","Learning Go","belajar-golang",{"title":60,"title_en":61,"slug":62},"Belajar Java","Learning Java","belajar-java",{"title":64,"title_en":65,"slug":66},"Belajar Pascal","Learning Pascal","belajar-pascal",{"collection":31,"item":68},{"__typename":31,"title":69,"title_en":69,"slug":70,"children_series":71},"Lab","lab",[],{"article":73,"renderedBody":705,"renderedTakeaway":706,"latestArticles":707,"toc":804},{"id":74,"title":75,"title_en":76,"slug":77,"excerpt":78,"excerpt_en":79,"has_translation":12,"published_at":80,"reading_time":14,"hidden_from_listing":15,"tags":81,"thumbnail_image":21,"og_image":88,"series":89,"authors":699,"body_locale":700,"available_locales":701,"body":703,"takeway":704,"prerequisite":21},112,"Aplikasimu lambat, dan kamu udah kepikiran ganti bahasa. Tunggu dulu","Your App Is Slow, and You're Already Thinking About Changing Languages. Wait.","aplikasimu-lambat-dan-kamu-udah-kepikiran-ganti-bahasa-tunggu-dulu","Ada endpoint yang grafiknya spike tiap beberapa menit, dan kepikiran buat pindah dari Go ke Rust? Discord dan Cloudflare beneran melakukannya — tapi cuma setelah kehabisan cara lain dan punya datanya. Yuk, bedah gejala mana yang beneran butuh rewrite, dan mana yang cuma butuh profiling. 🤔","An endpoint spikes every few minutes, and you are thinking of moving from Go to Rust? Discord and Cloudflare really did it, but only after running out of other options and having the data. Let's sort out which symptoms need a rewrite and which only need a profiler.","2026-10-08T09:40:00.000Z",[82,83,84,85,86,87],"backend","performance","golang","rust","system-design","profiling","\u002Fog\u002Faplikasimu-lambat-dan-kamu-udah-kepikiran-ganti-bahasa-tunggu-dulu.png",{"id":24,"title":25,"title_en":26,"slug":27,"description":21,"description_en":90,"articles":91,"children_series":692,"parent_series":21},"Everything that isn't a tutorial or a lab: explainers, opinions, and notes from building things.",[92,107,122,136,149,164,179,193,208,222,234,249,272,289,309,327,345,364,380,397,415,430,445,461,476,491,505,521,539,558,573,591,608,620,624,640,643,658,674],{"id":93,"title":94,"title_en":95,"slug":96,"excerpt":97,"excerpt_en":98,"has_translation":12,"published_at":99,"reading_time":14,"hidden_from_listing":15,"tags":100,"thumbnail_image":21,"og_image":104,"series":105,"authors":106},11,"Apa Kabar JavaScript? Rilis State of JS 2022","How's JavaScript Doing? The State of JS 2022 Results Are In","apa-kabar-javascript-rilis-state-of-js-2022","Survei State of JS 2022 resmi dirilis dengan lonjakan responden yang luar biasa! Framework dan build tool apa saja yang berhasil mencuri perhatian tahun ini?","The State of JS 2022 survey is out, and the respondent count jumped hard. Which frameworks and build tools actually stole the year?","2026-06-15T02:36:30.205Z",[101,102,103],"javascript","web","news","\u002Fog\u002Fapa-kabar-javascript-rilis-state-of-js-2022.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":108,"title":109,"title_en":110,"slug":111,"excerpt":112,"excerpt_en":113,"has_translation":12,"published_at":114,"reading_time":14,"hidden_from_listing":15,"tags":115,"thumbnail_image":21,"og_image":119,"series":120,"authors":121},12,"Buat PowerShell Lebih Berwarna dengan Starship","Make PowerShell More Colorful with Starship","buat-powershell-lebih-berwarna-dengan-starship","Bosen sama tampilan PowerShell yang hitam putih dan kaku? Yuk, sulap terminalmu jadi lebih interaktif, estetik, dan berwarna pakai Starship!","Tired of PowerShell looking black, white, and stiff? Turn your terminal into something interactive, aesthetic, and colorful with Starship!","2026-06-15T02:37:24.143Z",[116,117,118],"shell","windows","powershell","\u002Fog\u002Fbuat-powershell-lebih-berwarna-dengan-starship.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":123,"title":124,"title_en":125,"slug":126,"excerpt":127,"excerpt_en":128,"has_translation":12,"published_at":129,"reading_time":14,"hidden_from_listing":15,"tags":130,"thumbnail_image":21,"og_image":133,"series":134,"authors":135},13,"Hal yang Perlu Diperhatikan Sebelum Pindah ke Linux","Things to Consider Before Switching to Linux","hal-yang-perlu-diperhatikan-sebelum-pindah-ke-linux","Tertarik migrasi dari Windows ke Linux tapi masih ragu? Yuk, pelajari kelebihan, kekurangan, dan persiapan penting sebelum kamu benar-benar pindah!","Tempted to migrate from Windows to Linux but still unsure? Learn the upsides, the downsides, and the prep you need before you actually switch.","2026-06-15T02:37:24.570Z",[131,117,132],"linux","foss","\u002Fog\u002Fhal-yang-perlu-diperhatikan-sebelum-pindah-ke-linux.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":137,"title":138,"title_en":139,"slug":140,"excerpt":141,"excerpt_en":142,"has_translation":12,"published_at":143,"reading_time":14,"hidden_from_listing":15,"tags":144,"thumbnail_image":21,"og_image":146,"series":147,"authors":148},15,"Kenapa Ada Banyak Sekali Distro Linux?","Why Are There So Many Linux Distros?","kenapa-ada-banyak-sekali-distro-linux","Pernah bingung pas mau install Linux tapi malah disuguhi puluhan pilihan distro? Tenang, kalian nggak sendirian. Yuk, cari tahu kenapa distro Linux bisa sebanyak itu dan apa bedanya!","Were you confused when installing Linux handed you dozens of distro options? You're not alone. Here's why there are so many, and what actually differs.","2026-06-15T02:38:14.569Z",[131,145],"distro","\u002Fog\u002Fkenapa-ada-banyak-sekali-distro-linux.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":150,"title":151,"title_en":152,"slug":153,"excerpt":154,"excerpt_en":155,"has_translation":12,"published_at":156,"reading_time":14,"hidden_from_listing":15,"tags":157,"thumbnail_image":21,"og_image":161,"series":162,"authors":163},16,"Menggunakan TypeScript dengan Node.js dan Express","Using TypeScript with Node.js and Express","menggunakan-typescript-dengan-node-js-dan-express","Pengen bikin server Express di Node.js tapi pakai fitur-fitur keren TypeScript? Yuk, intip cara setup lengkapnya dari nol di sini!","Want to build an Express server on Node.js but use TypeScript's nicer features? Here's the complete setup from scratch.","2026-06-15T02:38:30.735Z",[101,158,159,160],"nodejs","express","typescript","\u002Fog\u002Fmenggunakan-typescript-dengan-node-js-dan-express.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":165,"title":166,"title_en":167,"slug":168,"excerpt":169,"excerpt_en":170,"has_translation":12,"published_at":171,"reading_time":14,"hidden_from_listing":15,"tags":172,"thumbnail_image":21,"og_image":176,"series":177,"authors":178},17,"Pemrograman Asynchronous vs Multithreading, Apa Sih Bedanya?","Asynchronous Programming vs Multithreading: What's the Real Difference?","pemrograman-asynchronous-dan-multithreading-apa-bedanya","Sering dengar istilah asynchronous dan multithreading pas ngoding tapi masih bingung bedanya? Yuk, kita bedah perbedaannya lewat analogi bikin mi instan yang gampang dipahami!","You keep hearing asynchronous and multithreading while coding, but the difference stays fuzzy? Let's take it apart with an instant-noodle analogy you can follow.","2026-06-15T02:38:31.093Z",[173,174,175],"pemrograman","concurrency","tips-coding","\u002Fog\u002Fpemrograman-asynchronous-dan-multithreading-apa-bedanya.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":180,"title":181,"title_en":182,"slug":183,"excerpt":184,"excerpt_en":185,"has_translation":12,"published_at":186,"reading_time":14,"hidden_from_listing":15,"tags":187,"thumbnail_image":21,"og_image":190,"series":191,"authors":192},18,"React, Angular, atau Vue: Pilih Framework yang Mana?","React, Angular, or Vue: Which Framework Should You Pick?","react-angular-vue-pilih-framework-yang-mana","Bingung menentukan pilihan di antara React, Angular, dan Vue untuk proyek web kalian berikutnya? Tenang, yuk kita bandingkan kelebihan masing-masing di sini!","Stuck choosing between React, Angular, and Vue for your next web project? Here's a side-by-side look at what each one actually does well.","2026-06-15T02:38:31.431Z",[188,189,18],"web-development","frontend","\u002Fog\u002Freact-angular-vue-pilih-framework-yang-mana.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":194,"title":195,"title_en":196,"slug":197,"excerpt":198,"excerpt_en":199,"has_translation":12,"published_at":200,"reading_time":14,"hidden_from_listing":15,"tags":201,"thumbnail_image":21,"og_image":205,"series":206,"authors":207},19,"Static Typing dan Dynamic Typing: Apa Bedanya?","Static Typing vs Dynamic Typing: What's the Difference?","static-typing-dan-dynamic-typing-apa-bedanya","Pernah bingung kenapa di Java tipe data harus dideklarasikan, sementara di Python tinggal pakai aja? Yuk, cari tahu perbedaan static typing dan dynamic typing di sini!","Ever wondered why Java makes you declare a type while Python just takes the value? Here's the difference between static typing and dynamic typing.","2026-06-15T02:39:15.869Z",[202,203,204],"programming","types","comparison","\u002Fog\u002Fstatic-typing-dan-dynamic-typing-apa-bedanya.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":209,"title":210,"title_en":211,"slug":212,"excerpt":213,"excerpt_en":214,"has_translation":12,"published_at":215,"reading_time":14,"hidden_from_listing":15,"tags":216,"thumbnail_image":21,"og_image":219,"series":220,"authors":221},20,"Tailwind v3.0 Rilis: Ada JIT, Arbitrary Values, dan Banyak Lagi!","Tailwind v3.0 Is Out: JIT, Arbitrary Values, and Plenty More","tailwind-v3-0-rilis-jit-arbitrary-dan-banyak-lagi","Tailwind CSS v3.0 resmi dirilis dengan segudang peningkatan performa yang bikin proses coding makin asyik. Apa aja sih fitur andalan yang wajib dicoba?","Tailwind CSS v3.0 shipped with a pile of performance upgrades that make styling more fun. Which features are worth trying first?","2026-06-15T02:39:25.372Z",[217,218,188],"tailwind-css","css-framework","\u002Fog\u002Ftailwind-v3-0-rilis-jit-arbitrary-dan-banyak-lagi.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":223,"title":224,"title_en":225,"slug":226,"excerpt":227,"excerpt_en":228,"has_translation":12,"published_at":229,"reading_time":14,"hidden_from_listing":15,"tags":230,"thumbnail_image":21,"og_image":231,"series":232,"authors":233},21,"TypeScript: JavaScript dengan Gaya dan Tipe Data Statis","TypeScript: JavaScript with Style and Static Types","typescript-javascript-dengan-gaya","Pernah pusing gara-gara runtime error di JavaScript? Yuk, kenalan dengan TypeScript yang bawa fitur static typing biar kode kalian makin rapi dan aman!","Tired of runtime errors in JavaScript? Meet TypeScript, which brings static typing so your code stays tidier and safer.","2026-06-15T02:39:25.712Z",[160,101,188],"\u002Fog\u002Ftypescript-javascript-dengan-gaya.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":235,"title":236,"title_en":237,"slug":238,"excerpt":239,"excerpt_en":240,"has_translation":12,"published_at":241,"reading_time":14,"hidden_from_listing":15,"tags":242,"thumbnail_image":21,"og_image":246,"series":247,"authors":248},22,"Yang Baru di Nuxt 3: Nitro Engine, Vite, dan Fitur Keren Lainnya","What's New in Nuxt 3: Nitro Engine, Vite, and Other Sharp Features","yang-baru-di-nuxt-3-nitro-vite-dan-banyak-lagi","Nuxt 3 hadir membawa perubahan arsitektur besar-besaran untuk menyelaraskan dengan Vue 3. Penasaran apa saja peningkatan performa dan fitur barunya?","Nuxt 3 brings a sweeping architecture change to line up with Vue 3. Here's what got faster and which new features are worth trying.","2026-06-15T02:39:26.086Z",[243,244,245],"nuxt-3","vue-3","frontend-framework","\u002Fog\u002Fyang-baru-di-nuxt-3-nitro-vite-dan-banyak-lagi.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":250,"title":251,"title_en":252,"slug":253,"excerpt":254,"excerpt_en":255,"has_translation":12,"published_at":256,"reading_time":257,"hidden_from_listing":15,"tags":258,"thumbnail_image":21,"og_image":263,"series":264,"authors":265},23,"Kenapa Pakai FastAPI dan Kenapa Dia Kenceng Banget?","Why Use FastAPI, and Why Is It So Fast?","kenapa-pakai-fastapi-dan-kenapa-dia-kenceng-banget","Sering dengar FastAPI tapi masih bingung kenapa dia bisa ngebut kayak NodeJS dan Go? Yuk, kita bedah bareng rahasia ASGI, sihir Pydantic, sampai dokumentasi otomatisnya!","You keep hearing about FastAPI but still wonder how it can keep pace with NodeJS and Go? Let's take apart ASGI, the Pydantic magic, and the automatic docs.","2026-08-16T13:22:46.956Z",4,[259,260,82,261,262],"fastapi","python","asgi","pydantic","\u002Fog\u002Fkenapa-pakai-fastapi-dan-kenapa-dia-kenceng-banget.png",{"id":24,"title":25,"title_en":26,"slug":27},[266],{"authors_id":267},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},2,"Inva","inva","Writer",{"id":273,"title":274,"title_en":275,"slug":276,"excerpt":277,"excerpt_en":278,"has_translation":12,"published_at":279,"reading_time":257,"hidden_from_listing":15,"tags":280,"thumbnail_image":21,"og_image":284,"series":285,"authors":286},24,"Kenapa Butuh Rate Limiting?","Why Do You Need Rate Limiting?","kenapa-butuh-rate-limiting","Pernah nggak sih endpoint login kalian diserbu bot yang nyoba nembak ribuan password cuma dalam semenit? Yuk, kita bahas kenapa Rate Limiting itu wajib, plus cara masangnya di FastAPI!","Has your login endpoint ever been hammered by a bot trying thousands of passwords in a single minute? Here is why rate limiting is mandatory, and how to wire it into FastAPI.","2026-08-17T04:00:27.728Z",[82,281,86,282,283,259],"rate-limiting","api","redis","\u002Fog\u002Fkenapa-butuh-rate-limiting.png",{"id":24,"title":25,"title_en":26,"slug":27},[287],{"authors_id":288},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":290,"title":291,"title_en":292,"slug":293,"excerpt":294,"excerpt_en":295,"has_translation":12,"published_at":296,"reading_time":14,"hidden_from_listing":15,"tags":297,"thumbnail_image":21,"og_image":301,"series":302,"authors":303},25,"Kenapa Event Queue Bikin Backend Kalian Nggak Gampang Tumbang?","Why an Event Queue Keeps Your Backend From Collapsing","kenapa-event-queue-bikin-backend-kamu-nggak-gampang-tumbang","Pernah klik tombol Beli Sekarang dan notifikasi suksesnya muncul dalam sekejap, padahal di belakang layar ada invoice, email, sampai push notification yang harus dibuat? Rahasianya Event Queue. Yuk, kita bedah!","You hit Buy Now and the success notice appears instantly, while an invoice, an email, and a push notification still have to be built behind the scenes. The secret is the event queue.","2026-08-17T04:11:23.685Z",[82,86,298,299,300],"event-queue","message-broker","architecture","\u002Fog\u002Fkenapa-event-queue-bikin-backend-kamu-nggak-gampang-tumbang.png",{"id":24,"title":25,"title_en":26,"slug":27},[304],{"authors_id":305},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},1,"Daffa Izzuddin","izzudd",{"id":310,"title":311,"title_en":312,"slug":313,"excerpt":314,"excerpt_en":315,"has_translation":12,"published_at":316,"reading_time":14,"hidden_from_listing":15,"tags":317,"thumbnail_image":21,"og_image":322,"series":323,"authors":324},26,"REST API vs gRPC: Kapan Sebaiknya Mulai Pindah dari JSON?","REST API vs gRPC: When Should You Start Moving Off JSON?","rest-api-vs-grpc-kapan-sebaiknya-pindah-dari-json","Pernah nggak sih sistem mikroservis kalian mulai kerasa berat padahal cuma tuker-tukeran data internal? Belum tentu masalah servernya. Yuk, kita bahas kapan REST + JSON masih juara dan kapan gRPC jadi jawabannya!","Has your microservice system started feeling heavy even though it only shuffles internal data? The server may not be the problem. Here is when REST + JSON still wins, and when gRPC is the answer.","2026-08-17T04:26:10.249Z",[82,318,319,320,321,86],"grpc","rest-api","microservices","protobuf","\u002Fog\u002Frest-api-vs-grpc-kapan-sebaiknya-pindah-dari-json.png",{"id":24,"title":25,"title_en":26,"slug":27},[325],{"authors_id":326},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":328,"title":329,"title_en":330,"slug":331,"excerpt":332,"excerpt_en":333,"has_translation":12,"published_at":334,"reading_time":14,"hidden_from_listing":15,"tags":335,"thumbnail_image":21,"og_image":340,"series":341,"authors":342},27,"Background Worker & Task Queue: Jangan Jalankan Proses Berat di Request HTTP!","Background Workers & Task Queues: Don't Run Heavy Work Inside an HTTP Request","background-worker-task-queue-jangan-jalankan-proses-berat-di-http","Pernah nggak sih kalian klik Daftar Akun lalu halaman loading muter sampai 15 detik tanpa kepastian? Penyebabnya klasik banget. Yuk, kita bahas kenapa proses berat nggak boleh dijalanin di request HTTP!","You tap Create Account and the loading page spins for 15 seconds with no answer? The cause is classic. Here is why heavy work does not belong in an HTTP request.","2026-08-17T04:26:10.284Z",[82,336,337,259,338,283,339],"task-queue","celery","rabbitmq","async","\u002Fog\u002Fbackground-worker-task-queue-jangan-jalankan-proses-berat-di-http.png",{"id":24,"title":25,"title_en":26,"slug":27},[343],{"authors_id":344},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"id":346,"title":347,"title_en":348,"slug":349,"excerpt":350,"excerpt_en":351,"has_translation":12,"published_at":352,"reading_time":14,"hidden_from_listing":15,"tags":353,"thumbnail_image":21,"og_image":359,"series":360,"authors":361},28,"Mengatasi N+1 Query Problem: Musuh Tersembunyi Performa Backend","Taming the N+1 Query Problem: The Hidden Enemy of Backend Performance","mengatasi-n-plus-1-query-problem-musuh-tersembunyi-backend","Pernah bikin endpoint yang cuma nampilin 50 postingan beserta nama penulisnya, tapi response-nya lambat banget? Bisa jadi kalian kena N+1 Query. Yuk, kita bedah penyebabnya dan cara memberantasnya!","You built an endpoint that only lists 50 posts with their author names, and the response takes 2-3 seconds? You may have hit N+1 Query. Here is the cause, and the cure.","2026-08-17T04:26:10.309Z",[354,355,356,357,358,83,82],"database","sql","orm","sqlalchemy","django","\u002Fog\u002Fmengatasi-n-plus-1-query-problem-musuh-tersembunyi-backend.png",{"id":24,"title":25,"title_en":26,"slug":27},[362],{"authors_id":363},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"id":365,"title":366,"title_en":367,"slug":368,"excerpt":369,"excerpt_en":370,"has_translation":12,"published_at":371,"reading_time":257,"hidden_from_listing":15,"tags":372,"thumbnail_image":21,"og_image":375,"series":376,"authors":377},29,"Memahami Database Indexing: Rahasia Query Super Cepat","Understanding Database Indexing: The Secret to Very Fast Queries","memahami-database-indexing-rahasia-query-super-cepat","Query kalian mulai lelet padahal data baru nembus jutaan baris? Sebelum buru-buru *upgrade* server, yuk cek dulu *index*-nya. Yuk, kita bedah gimana *database indexing* sebenernya kerja!","Your queries have gone sluggish even though the data only just crossed a million rows? Before you rush to upgrade the server, check the indexes. Here is how database indexing actually works.","2026-08-17T04:26:10.325Z",[354,355,373,374,83,82],"postgresql","indexing","\u002Fog\u002Fmemahami-database-indexing-rahasia-query-super-cepat.png",{"id":24,"title":25,"title_en":26,"slug":27},[378],{"authors_id":379},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"id":381,"title":382,"title_en":383,"slug":384,"excerpt":385,"excerpt_en":386,"has_translation":12,"published_at":387,"reading_time":257,"hidden_from_listing":15,"tags":388,"thumbnail_image":21,"og_image":392,"series":393,"authors":394},31,"Mengenal Redis: Lebih dari Sekadar In-Memory Cache","Getting to Know Redis: More Than Just an In-Memory Cache","mengenal-redis-lebih-dari-sekadar-in-memory-cache","Masih nganggep Redis cuma buat caching? Padahal dia bisa jadi leaderboard, message broker, sampai rate limiter lho. Yuk, kita bedah semua kemampuannya!","Still treating Redis as nothing but a cache? It can be a leaderboard, a message broker, even a rate limiter. Let's take apart everything it can do.","2026-08-17T04:26:10.360Z",[283,82,389,390,391,260],"cache","nosql","in-memory","\u002Fog\u002Fmengenal-redis-lebih-dari-sekadar-in-memory-cache.png",{"id":24,"title":25,"title_en":26,"slug":27},[395],{"authors_id":396},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"id":398,"title":399,"title_en":400,"slug":401,"excerpt":402,"excerpt_en":403,"has_translation":12,"published_at":404,"reading_time":257,"hidden_from_listing":15,"tags":405,"thumbnail_image":21,"og_image":410,"series":411,"authors":412},32,"Docker Compose untuk Developer Santai: Satu Perintah Buat Jalankan Semua Service","Docker Compose for the Relaxed Developer: One Command to Run Every Service","docker-compose-untuk-developer-santai","Pernah nggak kalian ngulang-ngulang perintah `docker run` cuma buat nyalain web, database, dan cache satu-satu? Capek, kan? Yuk, kita rapikan semuanya jadi satu file dan satu perintah!","Do you keep retyping `docker run` just to bring up the web app, the database, and the cache one at a time? Here is how to fold all of them into one file and one command.","2026-08-17T04:26:10.375Z",[406,407,408,409,82],"docker","docker-compose","devops","containers","\u002Fog\u002Fdocker-compose-untuk-developer-santai.png",{"id":24,"title":25,"title_en":26,"slug":27},[413],{"authors_id":414},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"id":416,"title":417,"title_en":418,"slug":419,"excerpt":420,"excerpt_en":421,"has_translation":12,"published_at":422,"reading_time":257,"hidden_from_listing":15,"tags":423,"thumbnail_image":21,"og_image":425,"series":426,"authors":427},33,"Berkenalan dengan RabbitMQ: Si Kurir Pesan Serba Bisa","Meet RabbitMQ: The Versatile Message Courier","berkenalan-dengan-rabbitmq-si-kurir-pesan-serba-bisa","Pernah nggak sih kalian klik Bayar Sekarang dan notifikasi suksesnya muncul seketika, padahal email konfirmasi baru nyampe beberapa detik kemudian? Rahasianya ada di satu kurir tak terlihat. Yuk, kita kenalan sama RabbitMQ!","You tap Pay Now and the success notice appears instantly, while the confirmation email only lands a few seconds later? The secret is one invisible courier. Meet RabbitMQ.","2026-08-17T04:26:10.388Z",[338,299,82,320,339,424],"queue","\u002Fog\u002Fberkenalan-dengan-rabbitmq-si-kurir-pesan-serba-bisa.png",{"id":24,"title":25,"title_en":26,"slug":27},[428],{"authors_id":429},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":431,"title":432,"title_en":433,"slug":434,"excerpt":435,"excerpt_en":436,"has_translation":12,"published_at":437,"reading_time":257,"hidden_from_listing":15,"tags":438,"thumbnail_image":21,"og_image":440,"series":441,"authors":442},34,"Kenapa Backend Butuh Database Connection Pooling?","Why Does a Backend Need Database Connection Pooling?","kenapa-backend-butuh-database-connection-pooling","Pernah kena error Too many connections padahal CPU server masih santai di angka 30%? Masalahnya bukan di query kalian, tapi di cara aplikasi ngobrol sama database. Yuk, kenalan sama Connection Pooling!","Ever hit Too many connections while server CPU is still lounging at 30%? The problem is not your queries, it is how the application talks to the database. Meet connection pooling.","2026-08-24T06:06:47.088Z",[82,354,373,439,86,83],"connection-pooling","\u002Fog\u002Fkenapa-backend-butuh-database-connection-pooling.png",{"id":24,"title":25,"title_en":26,"slug":27},[443],{"authors_id":444},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":446,"title":447,"title_en":448,"slug":449,"excerpt":450,"excerpt_en":451,"has_translation":12,"published_at":452,"reading_time":14,"hidden_from_listing":15,"tags":453,"thumbnail_image":21,"og_image":456,"series":457,"authors":458},35,"Mengenal Idempotency Key: Rahasia Anti-Double Charge di API Pembayaran","Idempotency Keys: The Secret to No Double Charges in Payment APIs","mengenal-idempotency-key-rahasia-anti-double-charge-di-api-pembayaran","Pernah internet kalian ngadat pas lagi nekan tombol Bayar, terus panik ngekliknya tiga kali? Kalau endpoint-nya nggak idempotent, saldo bisa kepotong berkali-kali. Yuk, kenalan sama Idempotency Key!","Ever had your connection stall right as you hit Pay, then panic-clicked three times? If the endpoint is not idempotent, the balance can be cut more than once. Meet the idempotency key.","2026-08-24T06:09:11.645Z",[82,282,454,86,455,259,283],"idempotency","payment","\u002Fog\u002Fmengenal-idempotency-key-rahasia-anti-double-charge-di-api-pembayaran.png",{"id":24,"title":25,"title_en":26,"slug":27},[459],{"authors_id":460},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":462,"title":463,"title_en":464,"slug":465,"excerpt":466,"excerpt_en":467,"has_translation":12,"published_at":468,"reading_time":14,"hidden_from_listing":15,"tags":469,"thumbnail_image":21,"og_image":471,"series":472,"authors":473},36,"Mencegah Efek Domino di Mikroservis dengan Circuit Breaker Pattern","Stopping the Domino Effect in Microservices With the Circuit Breaker Pattern","mencegah-efek-domino-di-mikroservis-dengan-circuit-breaker-pattern","Pernah nggak sih, satu fitur minor di aplikasi kalian mendadak lelet, eh malah bikin halaman checkout dan login ikut down total? Itu namanya *cascading failure*. Yuk, kita bahas gimana *circuit breaker pattern* (pemutus arus) nyelametin arsitektur kalian!","Ever had one minor feature go slow and take the checkout page and login down with it? That is a cascading failure. Here is how the circuit breaker pattern saves your architecture.","2026-08-24T06:13:36.734Z",[82,320,470,86,300],"circuit-breaker","\u002Fog\u002Fmencegah-efek-domino-di-mikroservis-dengan-circuit-breaker-pattern.png",{"id":24,"title":25,"title_en":26,"slug":27},[474],{"authors_id":475},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":477,"title":478,"title_en":479,"slug":480,"excerpt":481,"excerpt_en":482,"has_translation":12,"published_at":483,"reading_time":14,"hidden_from_listing":15,"tags":484,"thumbnail_image":21,"og_image":486,"series":487,"authors":488},37,"Kenapa Backend Butuh Graceful Shutdown? Jangan Asal Cabut Kabel Server!","Why Does a Backend Need Graceful Shutdown? Don't Yank the Server Cable","kenapa-backend-butuh-graceful-shutdown-jangan-asal-cabut-kabel-server","Pernah nggak sih, abis kalian nge-*deploy* backend, beberapa detik kemudian monitoring malah banjir *alert* **502 Bad Gateway** dari pengguna? Saldo udah kepotong tapi status order belum keburu ter-*update* di *database*. Yuk, kita bahas kenapa *graceful shutdown* itu wajib banget!","You deploy the backend and seconds later monitoring floods with 502 Bad Gateway alerts. The balance was cut but the order status never reached the database. Here is why graceful shutdown is mandatory.","2026-08-24T06:15:59.150Z",[82,408,259,406,485,86,300],"kubernetes","\u002Fog\u002Fkenapa-backend-butuh-graceful-shutdown-jangan-asal-cabut-kabel-server.png",{"id":24,"title":25,"title_en":26,"slug":27},[489],{"authors_id":490},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":492,"title":493,"title_en":494,"slug":495,"excerpt":496,"excerpt_en":497,"has_translation":12,"published_at":498,"reading_time":14,"hidden_from_listing":15,"tags":499,"thumbnail_image":21,"og_image":500,"series":501,"authors":502},38,"Optimistic vs Pessimistic Locking: Mengatasi Rebutan Data di Database","Optimistic vs Pessimistic Locking: Handling the Data Fight in a Database","optimistic-vs-pessimistic-locking-mengatasi-rebutan-data-di-database","Seribu orang ngeklik \"Beli Sekarang\" di detik yang sama buat rebutan 1 tiket terakhir. Kira-kira apa yang bakal terjadi di server kalian? Yuk, kita banding *pessimistic* sama *optimistic locking*!","A thousand people tap Buy Now in the same second, fighting over the last ticket. What happens on your server? Here is pessimistic against optimistic locking.","2026-08-24T06:16:37.198Z",[82,354,174,373,86,300,355],"\u002Fog\u002Foptimistic-vs-pessimistic-locking-mengatasi-rebutan-data-di-database.png",{"id":24,"title":25,"title_en":26,"slug":27},[503],{"authors_id":504},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":506,"title":507,"title_en":508,"slug":509,"excerpt":510,"excerpt_en":511,"has_translation":12,"published_at":512,"reading_time":14,"hidden_from_listing":15,"tags":513,"thumbnail_image":21,"og_image":516,"series":517,"authors":518},39,"Database Sharding vs Partitioning: Kapan Data Perlu Dipecah?","Database Sharding vs Partitioning: When Does Data Need Splitting?","database-sharding-vs-partitioning-kapan-data-perlu-dipecah","Tabel transaksi kalian udah nembus miliaran baris dan tiap *query* laporan bikin *CPU* melonjak 100%? Sebelum asal pecah data, yuk kita pahami dulu beda *sharding* sama *partitioning*!","Your transaction table has crossed a billion rows and every report query spikes CPU to 100%? Before splitting data blindly, understand the difference between sharding and partitioning.","2026-08-24T06:17:10.104Z",[82,354,373,86,300,514,515],"sharding","scaling","\u002Fog\u002Fdatabase-sharding-vs-partitioning-kapan-data-perlu-dipecah.png",{"id":24,"title":25,"title_en":26,"slug":27},[519],{"authors_id":520},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":522,"title":523,"title_en":524,"slug":525,"excerpt":526,"excerpt_en":527,"has_translation":12,"published_at":528,"reading_time":14,"hidden_from_listing":15,"tags":529,"thumbnail_image":21,"og_image":534,"series":535,"authors":536},40,"Menyelami Distributed Tracing: Melacak Jejak Request di Balik Keruwetan Microservices","Inside Distributed Tracing: Following a Request Through Microservice Chaos","menyelami-distributed-tracing-melacak-jejak-request-microservices","Satu klik Checkout di backend kalian bisa memicu puluhan network call antar-service, tapi pas ada yang lambat, service mana sebenernya yang salah? Log aja nggak cukup buat njawab itu. Yuk, kita bedah Distributed Tracing!","One Checkout click in your backend can trigger dozens of network calls between services, but when one is slow, which service is at fault? Logs alone cannot answer that. Here is distributed tracing.","2026-08-31T02:33:31.602Z",[82,530,531,532,320,300,533],"distributed-systems","observability","open-telemetry","tracing","\u002Fog\u002Fmenyelami-distributed-tracing-melacak-jejak-request-microservices.png",{"id":24,"title":25,"title_en":26,"slug":27},[537],{"authors_id":538},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":540,"title":541,"title_en":542,"slug":543,"excerpt":544,"excerpt_en":545,"has_translation":12,"published_at":546,"reading_time":14,"hidden_from_listing":15,"tags":547,"thumbnail_image":21,"og_image":553,"series":554,"authors":555},41,"Log Aggregation Modern dengan Grafana Loki dan Vector: Hemat Storage Tanpa Indeks Raksasa","Modern Log Aggregation With Grafana Loki and Vector: Save Storage Without a Giant Index","log-aggregation-modern-grafana-loki-vector-hemat-storage","Klaster Elasticsearch kalian rakus RAM dan boros storage, ya? Padahal nggak semua kata di log perlu diindeks, lho. Yuk, kita kenalan sama duet Grafana Loki dan Vector yang jauh lebih ramah kantong!","Your Elasticsearch cluster is greedy for RAM and wasteful with storage? Not every word in a log needs indexing. Meet the duo of Grafana Loki and Vector, far kinder to your wallet.","2026-09-01T15:45:01.066Z",[82,548,549,550,551,552,408,86],"observa","logging","grafana","loki","vector","\u002Fog\u002Flog-aggregation-modern-grafana-loki-vector-hemat-storage.png",{"id":24,"title":25,"title_en":26,"slug":27},[556],{"authors_id":557},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":559,"title":560,"title_en":560,"slug":561,"excerpt":562,"excerpt_en":563,"has_translation":12,"published_at":564,"reading_time":14,"hidden_from_listing":15,"tags":565,"thumbnail_image":21,"og_image":568,"series":569,"authors":570},45,"Rate Limiting Algorithms: Token Bucket vs Leaky Bucket vs Sliding Window","rate-limiting-algorithms-token-bucket-leaky-bucket-sliding-window","Udah tahu kenapa *rate limiting* itu wajib, tapi masih bingung beda Token Bucket, Leaky Bucket, Fixed Window, dan Sliding Window? Yuk, kita adu empat algoritma ini bareng-bareng dan cari tahu mana yang paling pas buat API kalian!","You already know why *rate limiting* is mandatory, but the difference between Token Bucket, Leaky Bucket, Fixed Window, and Sliding Window is still fuzzy. Let's put all four algorithms side by side and find the one that fits your API.","2026-09-09T01:57:55.886Z",[82,86,281,283,566,300,567],"api-gateway","traffict-management","\u002Fog\u002Frate-limiting-algorithms-token-bucket-leaky-bucket-sliding-window.png",{"id":24,"title":25,"title_en":26,"slug":27},[571],{"authors_id":572},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":574,"title":575,"title_en":576,"slug":577,"excerpt":578,"excerpt_en":579,"has_translation":12,"published_at":580,"reading_time":14,"hidden_from_listing":15,"tags":581,"thumbnail_image":21,"og_image":586,"series":587,"authors":588},46,"Mengapa C Mulai Ditinggalkan: Duel Filosofi Keamanan Rust vs Kesederhanaan Zig","Why C Is Being Left Behind: Rust's Safety Philosophy vs Zig's Simplicity","mengapa-c-mulai-ditinggalkan-duel-filosofi-rust-vs-zig","Pernah denger kalau sekitar 70% celah keamanan kritis software modern itu asalnya dari bahasa C? Iya, bahasa yang udah jadi fondasi internet ini 😅. Yuk, kita bedah duel filosofi Rust vs Zig dan cari tahu mana yang pas buat proyek kalian!","Have you heard that around 70% of critical security holes in modern software come from C? Yes, the language the internet is built on. Let's unpack the philosophy duel between Rust and Zig and see which one fits your project.","2026-09-09T02:19:05.266Z",[582,85,583,584,585],"system-programming","zig","c","memory-safety","\u002Fog\u002Fmengapa-c-mulai-ditinggalkan-duel-filosofi-rust-vs-zig.png",{"id":24,"title":25,"title_en":26,"slug":27},[589],{"authors_id":590},{"id":268,"name":269,"slug":270,"role":271,"profile_picture":21},{"id":592,"title":593,"title_en":594,"slug":595,"excerpt":596,"excerpt_en":597,"has_translation":12,"published_at":598,"reading_time":599,"hidden_from_listing":15,"tags":600,"thumbnail_image":21,"og_image":605,"series":606,"authors":607},107,"Cara kamu pakai AI lebih menentukan daripada apakah kamu pakai AI","How You Use AI Matters More Than Whether You Use AI","cara-pakai-ai-menentukan-pemahaman-kode","Fiturnya selesai, test-nya hijau, dan dua hari kemudian kamu baca kodenya kayak kode orang lain. Ternyata ada satu cara pakai AI yang bikin kamu paham lebih baik daripada nggak pakai AI sama sekali. Yuk lihat datanya.","The feature is done, the tests are green, and two days later you read the code like a stranger wrote it. It turns out one way of using AI leaves you understanding more than not using AI at all. Here is the data.","2026-10-10T11:08:29.000Z",7,[601,602,603,604],"ai","belajar","kognitif","ai-assisted-coding","\u002Fog\u002Fcara-pakai-ai-menentukan-pemahaman-kode.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":609,"title":610,"title_en":611,"slug":612,"excerpt":613,"excerpt_en":614,"has_translation":12,"published_at":13,"reading_time":14,"hidden_from_listing":15,"tags":615,"thumbnail_image":21,"og_image":617,"series":618,"authors":619},109,"State of JS 2025 Bilang Developer Cuma Pakai 2,6 Framework Sepanjang Karier","State of JS 2025 Says Developers Use Only 2.6 Frameworks in a Whole Career","state-of-js-2025-developer-cuma-pakai-2-6-framework","Coba hitung semua framework yang pernah kamu coba, lalu potong jadi yang beneran dipakai buat kerjaan. Daftarnya jauh lebih pendek dari dugaan. State of JS 2025 nanya ke 13.002 developer, dan rata-ratanya cuma 2,6. Yuk, lihat angkanya bareng-bareng!","Try counting every framework you have ever touched, then cut the list down to what you actually used for work. It comes out far shorter than you would guess. State of JS 2025 asked 13,002 developers, and the average is only 2.6.",[101,616,18,17],"survei","\u002Fog\u002Fstate-of-js-2025-developer-cuma-pakai-2-6-framework.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":6,"title":7,"title_en":8,"slug":9,"excerpt":10,"excerpt_en":11,"has_translation":12,"published_at":13,"reading_time":14,"hidden_from_listing":15,"tags":621,"thumbnail_image":21,"og_image":22,"series":622,"authors":623},[17,18,19,20],{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":625,"title":626,"title_en":627,"slug":628,"excerpt":629,"excerpt_en":630,"has_translation":12,"published_at":631,"reading_time":632,"hidden_from_listing":15,"tags":633,"thumbnail_image":21,"og_image":635,"series":636,"authors":637},111,"Rust Masuk Kernel Linux: Revolusi Nyata atau Cuma Euforia RIIR?","Rust Enters the Linux Kernel: A Real Revolution or Just RIIR Euphoria?","rust-di-kernel-linux-revolusi-atau-cuma-hype","Selama lebih dari 30 tahun kernel Linux cuma mau bahasa C, lalu tiba-tiba Rust resmi masuk di Linux 6.1. Itu revolusi beneran atau cuma kebawa euforia meme Rewrite It In Rust? Yuk lihat datanya, bukan cuma opininya.","For more than 30 years the Linux kernel only wanted C, then Rust officially landed in Linux 6.1. Is that a real revolution, or just the pull of the Rewrite It In Rust meme? Let's look at the data, not the opinions.","2026-10-07T03:40:24.000Z",6,[85,131,634,582,585],"kernel","\u002Fog\u002Frust-di-kernel-linux-revolusi-atau-cuma-hype.png",{"id":24,"title":25,"title_en":26,"slug":27},[638],{"authors_id":639},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"id":74,"title":75,"title_en":76,"slug":77,"excerpt":78,"excerpt_en":79,"has_translation":12,"published_at":80,"reading_time":14,"hidden_from_listing":15,"tags":81,"thumbnail_image":21,"og_image":88,"series":641,"authors":642},{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":644,"title":645,"title_en":646,"slug":647,"excerpt":648,"excerpt_en":649,"has_translation":12,"published_at":650,"reading_time":599,"hidden_from_listing":15,"tags":651,"thumbnail_image":21,"og_image":655,"series":656,"authors":657},113,"Garbage Collector dituduh bikin aplikasi lemot. Seberapa adil tuduhan itu?","The Garbage Collector Gets Blamed for Slow Apps. How Fair Is That?","garbage-collector-dituduh-bikin-aplikasi-lemot","GC sering jadi tersangka utama tiap aplikasi melambat, dan banyak yang percaya dia jalan pakai timer tiap beberapa detik. Dua-duanya nggak sepenuhnya benar. Yuk bedah cara kerja GC dari dalam, plus satu pengecualian yang bikin Discord kesakitan tiap 2 menit. 🤔","GC is always the prime suspect when an app slows down, and many people believe it runs on a timer every few seconds. Both beliefs are only partly right. Let's open up how GC works inside, plus one exception that made Discord hurt every 2 minutes.","2026-10-09T01:34:00.000Z",[82,83,84,652,653,654],"jvm","garbage-collector","memory","\u002Fog\u002Fgarbage-collector-dituduh-bikin-aplikasi-lemot.png",{"id":24,"title":25,"title_en":26,"slug":27},[],{"id":659,"title":660,"title_en":661,"slug":662,"excerpt":663,"excerpt_en":664,"has_translation":12,"published_at":665,"reading_time":14,"hidden_from_listing":15,"tags":666,"thumbnail_image":21,"og_image":669,"series":670,"authors":671},118,"Belajar Idempotency dari Stok yang Keburu Kebeli","Learning Idempotency from Stock That Got Bought Too Soon","pub-sub-sendiri-lalu-belajar-idempotency","Tugas pertama saya di kerjaan pertama: sistem yang kewalahan menerima webhook. Saya putuskan membangun pub\u002Fsub sendiri dari nol, dan pelajaran yang akhirnya datang bukan dari situ — tapi dari stok gudang yang keburu kebeli orang.","My first task at my first job was a system drowning in webhooks. I decided to build my own pub\u002Fsub from scratch, and the lesson that finally landed came from somewhere else: warehouse stock that got bought before the count caught up.","2026-10-10T14:03:00.000Z",[454,667,337,338,668],"event-driven","post-mortem","\u002Fog\u002Fpub-sub-sendiri-lalu-belajar-idempotency.png",{"id":24,"title":25,"title_en":26,"slug":27},[672],{"authors_id":673},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"id":675,"title":676,"title_en":677,"slug":678,"excerpt":679,"excerpt_en":680,"has_translation":12,"published_at":681,"reading_time":632,"hidden_from_listing":15,"tags":682,"thumbnail_image":21,"og_image":685,"series":686,"authors":687},124,"Rahasia Zero-Cost Abstraction di Rust, Dibedah dari Assembly","The Secret of Zero-Cost Abstractions in Rust, Opened Up from the Assembly","rahasia-zero-cost-abstraction-di-rust","\"Abstraksi tinggi selalu lebih lambat\" itu anggapan lama. Iterator Rust bisa menyamai loop manual, dan assembly-nya bisa kita baca sendiri. Yuk, kita bongkar!","\"Higher abstractions are always slower\" is an old assumption. Rust iterators can match a hand-written loop, and the assembly is there for you to read. Let's take it apart.","2026-10-13T05:19:00.000Z",[85,83,582,683,684],"compiler","optimasi","\u002Fog\u002Frahasia-zero-cost-abstraction-di-rust.png",{"id":24,"title":25,"title_en":26,"slug":27},[688,690],{"authors_id":689},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"authors_id":691},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},[693,695,697],{"id":694,"title":35,"slug":37,"description":21,"cover_image":21},8,{"id":108,"title":39,"slug":41,"description":696,"cover_image":21},"Mengenal konsep dasar pemrograman. Pada seri ini kita akan membahas tentang bagaimana cara berkutat dengan variabel, mengatur alur program dengan struktur kontrol, me-reuse kode dengan fungsi, dan aturan sintaks pada bahasa pemrograman. Konsep-konsep ini pastinya akan sangat berguna ketika kita mulai mengkoding ceria.",{"id":137,"title":43,"slug":45,"description":698,"cover_image":21},"Algoritma dan struktur data adalah konsep fundamental dalam dunia pemrograman. Keduanya sangat penting dalam menyelesaikan masalah yang kompleks dengan efisien dan efektif. Dalam seri ini kita akan membahas berbagai macam struktur data dan algoritma yang harapannya bisa bikin kode kita lebih efektif dan efisien. Yuk belajar ASD bareng Inva! 👋",[],"id",[702,700],"en","Bayangin ada satu *endpoint* yang grafik latency-nya nggak rata. Kebanyakan waktu dia adem, tapi tiba-tiba naik lancip, turun lagi, lalu naik lagi. Kamu mandangin grafik itu sambil ngopi, dan yang muncul di kepala **bukan profiler**. Yang muncul adalah kalimat, \"kayaknya pindah bahasa aja deh.\" 😅\n\nDorongan itu wajar banget, kok. Ada perusahaan besar yang beneran pindah bahasa dan hasilnya bagus. Tapi ada yang jarang kita perhatiin: dua cerita paling terkenal itu **nggak dimulai dari bahasa baru**. Mereka mulai dari angka, dari ngukur sampai kelihatan persis di mana yang mentok. Joel Spolsky pernah bilang bahwa [menulis ulang dari nol adalah \"the single worst strategic mistake that any software company can make\"](https:\u002F\u002Fwww.joelonsoftware.com\u002F2000\u002F04\u002F06\u002Fthings-you-should-never-do-part-i\u002F), dan kita bakal lihat kenapa kalimat itu masih relevan.\n\nYuk kita bedah tiga kasus nyata, plus satu kontra-contoh. Tujuannya bukan bilang pindah bahasa itu salah. Tujuannya biar kamu tahu kapan solusi murah udah habis, dan kapan baru boleh mikir *rewrite*.\n\n## Kenapa *Discord* akhirnya pindah dari Go ke Rust?\n\nYang mereka pindahkan namanya Read States. Tugasnya cuma nyatet channel dan pesan mana yang udah kamu baca. Sederhana, tapi layanan ini dipanggil tiap kali kita connect, tiap kali kirim pesan, dan tiap kali buka pesan. Jadi dia nempel di hampir semua hal yang kita lakuin di sana, lho.\n\nSkalanya gede, nih. Ada [miliaran Read State](https:\u002F\u002Fdiscord.com\u002Fblog\u002Fwhy-discord-is-switching-from-go-to-rust), satu per user per channel. Tiap instance cache memegang jutaan user dan puluhan juta Read State. Per detiknya ada ratusan ribu update cache dan puluhan ribu write *database*.\n\nGejalanya emang aneh tapi konsisten: spike latency dan CPU tiap sekitar 2 menit. Penyebabnya, Go memaksa *GC* jalan minimal tiap 2 menit, terlepas dari seberapa besar *heap*-nya tumbuh. Dari [grafik di blog mereka](https:\u002F\u002Fdiscord.com\u002Fblog\u002Fwhy-discord-is-switching-from-go-to-rust), CPU naik dari sekitar 20% ke 30–40%. *Average response time* naik dari sekitar 0–2 ms ke sekitar 25 ms. Yang p95 melonjak dari sekitar 0–10 ms ke sekitar 300 ms. Catat ya, itu p95, bukan p99, dan semua angka ini hasil baca grafik, jadi sifatnya perkiraan.\n\nNah, bagian pentingnya: mereka nggak langsung pindah. Mereka nyoba tuning dulu, dan dua-duanya gagal. Pertama, mengubah GC percent dari *endpoint*. Hasilnya nggak ngaruh, soalnya alokasi mereka nggak cukup cepat untuk memicu GC lebih sering. Kedua, memperkecil *LRU cache*. Spike-nya memang mengecil, tapi p99-nya naik karena cache jadi lebih sering *miss* dan bebannya pindah ke *database*.\n\nJadi mereka udah ketemu tembok, tuh. Bukan tembok malas, tapi tembok yang diukur dan dicoba dobrak dua kali. Baru setelah itu mereka porting ke Rust, dan kata mereka latency-nya sama bagusnya dengan versi Go, dan **nggak ada spike sama sekali**. Dengan optimasi dasar aja versi Rust sudah mengalahkan versi Go yang di-tuning mati-matian. Setelah diprofil lagi, Rust menang di latency, CPU, dan memori.\n\nBeberapa hari kemudian kapasitas cache dinaikkan ke 8 juta Read State, hal yang nggak mungkin di versi Go karena GC-nya makin lama makin berat. Hasilnya, *average* response time jadi satuan mikrodetik, dan Max @mention jadi satuan milidetik. Tapi lihat catatan kaki mereka sendiri: mereka nggak berpikir kamu harus menulis ulang semuanya di Rust cuma karena bisa.\n\n## Cloudflare: kodenya lebih cepat, atau arsitekturnya?\n\nCerita kedua datang dari Cloudflare, yang mengganti *NGINX* dengan *Pingora*. Menurut [tulisan mereka soal Pingora](https:\u002F\u002Fblog.cloudflare.com\u002Fhow-we-built-pingora-the-proxy-that-connects-cloudflare-to-the-internet\u002F), sistem barunya melayani lebih dari 1 triliun *request* per hari. Dia cuma butuh sepertiga CPU dan memori dibanding infrastruktur proxy sebelumnya. Di produksi, hematnya 70% CPU dan 67% memori di beban yang sama. TTFB juga turun 5 ms di median dan 80 ms di p95.\n\nSekarang bagian yang paling sering kelewat. Hemat itu **bukan karena kodenya lebih cepat**, lho. Layanan lama mereka juga udah sub-milidetik. Yang lebih cepat itu cara dia berbagi koneksi. 🤯\n\nMasalahnya ada di arsitektur NGINX, sih. Satu *request* cuma dilayani satu *worker*, yaitu satu proses OS. Load-nya nggak seimbang antar core, dan *connection pool*-nya per *worker*. Akibatnya makin banyak *worker* ditambah, makin jelek rasio *reuse* koneksinya. Ini kayak kalau tiap orang punya galon air sendiri di mejanya. Kamu butuh minum, galon sebelah masih penuh, tapi nggak bisa diambil, jadi kamu beli galon baru lagi. Tambah orang bukan bikin galon yang ada lebih kepakai, malah nambah galon baru terus.\n\nPingora dibangun dengan multithreading, *runtime* async *Tokio*, dan *work stealing*. Mereka juga bikin HTTP *library* sendiri karena *traffic* di internet banyak yang nggak patuh RFC. Hasilnya, koneksi baru yang dibuat per detik tinggal sepertiga. Untuk satu pelanggan besar, rasio *reuse* koneksi naik dari 87,1% ke 99,92%, dan koneksi baru ke origin turun 160x. Terjemahan blog mereka: \"kami menghemat 434 tahun waktu *handshake* setiap hari.\"\n\nAda alasan lain mereka pindah. C nggak *memory safe*, Lua kurang performan dan nggak ada *static typing*, dan komunitas NGINX bergerak lambat. Setelah melayani beberapa ratus triliun *request*, Pingora belum pernah crash karena kode service-nya. Sistem ini lalu [di-open source-kan pada 28 Februari 2024](https:\u002F\u002Fblog.cloudflare.com\u002Fpingora-open-source).\n\n## Jadi sebenarnya apa yang kamu beli dengan pindah bahasa?\n\nDua kemenangan tadi datang dari dua mekanisme yang beda. Di *Discord*, yang hilang adalah GC. Di Cloudflare, yang didapat adalah kemampuan berbagi resource lintas *thread*. Itu aja, sih. Bukan \"Rust lebih cepat\" secara magis, tapi ada **satu batasan spesifik** yang akhirnya lenyap.\n\nKalau masalahmu bukan salah satu dari dua itu, ganti bahasa nggak akan mengubah apa pun. Query lambat tetap lambat di Rust. N+1 tetap N+1 di Go. Kamu cuma bakal punya masalah yang sama dengan sintaks yang berbeda, plus biaya belajar yang nggak kecil, lho.\n\nKutipan Knuth sering dipotong jadi slogan. Versi lengkapnya begini: *\"We should forget about small efficiencies, say about 97% of the time: premature optimization is the root of all evil. Yet we should not pass up our opportunities in that critical 3%.\"* Slogan itu sering dipakai buat menghindari ngukur, padahal Knuth justru minta kita tahu kode mana yang masuk 3% itu. Dan satu-satunya cara tahu adalah dengan *profiler*.\n\n## Gimana kalau masalahmu memang GC, tapi bahasanya Python?\n\nDi sini ada kabar yang menenangkan. Kalau masalahmu memang GC, kamu nggak selalu butuh bahasa baru. Pada Januari 2017, [Instagram mematikan GC di Python-nya](https:\u002F\u002Finstagram-engineering.com\u002Fdismissing-python-garbage-collection-at-instagram-4dca40b29172) dan langsung jadi 10% lebih efisien. Nggak pindah bahasa, nggak ganti *framework*, **cuma mematikan satu fitur**. 🙌\n\nKonteksnya gini. Instagram jalan dengan *Django* multi-proses, pakai *uWSGI* model *pre-fork*. Satu master mem-*fork* dirinya jadi puluhan *worker* yang melayani *request*. Idealnya, memori yang dibagi antar *worker* itu tetap dibagi.\n\nTapi temuannya lain. *Shared memory* tiap *worker* turun dari sekitar 250 MB ke sekitar 140 MB, kira-kira sepertiga, cuma dalam beberapa detik setelah *worker* lahir. Sebabnya, Python pakai *reference counting*. Setiap kali sebuah objek dibaca, counter-nya ditulis. Tulisan itu memicu *copy-on-write* di kernel, jadi halaman memori yang tadinya dibagi berubah jadi milik masing-masing proses.\n\nCoba pikirin sebentar, deh. Gejala \"GC bikin latency jelek\" muncul di tiga tempat berbeda: di Go (*Discord*), di JVM Cassandra (*Discord* lagi), dan di Python (Instagram). Jawabannya beda-beda. *Discord* pindah bahasa, Instagram cuma ganti konfigurasi. Yang salah bukan bahasanya, tapi keputusan yang diambil di sekitar bahasanya.\n\n## Kadang bottleneck-nya bahkan bukan bahasa\n\nIni kontra-contohnya, dan lagi-lagi dari *Discord*. Pada 2023, mereka pindah dari *Cassandra* ke *ScyllaDB*, dan [177 node jadi 72 node](https:\u002F\u002Fdiscord.com\u002Fblog\u002Fhow-discord-stores-trillions-of-messages). P99 untuk mengambil pesan historis turun dari 40–125 ms jadi 15 ms. Insert pesan turun dari 5–70 ms jadi 5 ms.\n\nAkar masalah aslinya adalah *hot partition*, konkurensi yang nggak terkendali, dan GC di JVM Cassandra. Yang menyelesaikannya adalah *request coalescing* plus *consistent hash routing* (ditulis di Rust), serta *database* yang bebas GC. Jadi bahkan di cerita \"pindah ke Rust\", sebagian besar kemenangannya datang dari desain dan pilihan komponen yang tepat.\n\nBonusnya, migrator-nya ditulis di Rust dan estimasi migrasinya berubah dari 3 bulan jadi 9 hari. Puncaknya jalan di 3,2 juta pesan per detik. Jadi Rust tetap punya tempat yang jelas, tapi tempatnya sempit dan spesifik.\n\n## Kapan *rewrite* boleh dipertimbangkan?\n\nAlasan Spolsky masuk akal: kode lama itu sudah dipakai, sudah diuji, dan ratusan *bug* sudah ditemukan dan diperbaiki. Menulis ulang berarti membuang semua pengetahuan itu. Dia juga bilang soal performa, \"1% of the work gets you 99% of the bang.\" Kalau 1% itu belum kamu cari, jangan loncat ke 100%.\n\nJadi rewrite itu **keputusan terakhir**, bukan langkah pertama. Ada tiga pertanyaan yang harus kamu punya datanya.\n\n- Udah habis belum usaha yang lebih murah, seperti *query*, indeks, *caching*, N+1, dan *connection pool*? Ingat, Cloudflare mengevaluasi pilihannya tiap kuartal selama beberapa tahun.\n- Bottleneck-nya kebukti berasal dari *runtime* atau arsitektur, bukan dari kode kita sendiri?\n- Komponennya cukup terisolasi? *Discord* cuma memindahkan satu service, dan Cloudflare cuma mengganti layer proxy.\n\nKalau belum bisa jawab ketiganya dengan data, jawabannya sekarang adalah belum. Dan itu jawaban yang sehat, kok, bukan tanda kamu pengecut.\n\n## Kalau tetap harus pindah, pilih yang mana?\n\nIni cuma pendapat, ya. Pilih Go kalau bebanmu didominasi I\u002FO dan kamu butuh cepat produktif. Pilih Rust kalau bebanmu CPU-bound, kamu memegang cache besar di RAM, kamu bangun infrastruktur jaringan atau storage, atau kamu butuh p99 yang datar tanpa toleransi jeda GC.\n\nDan kalau Python atau Node-mu masih jalan baik setelah semua usaha murah dicoba, nggak ada yang memaksa kamu pindah. Bahasa yang dianggap lambat pun bukan masalah kalau dipakai dengan benar.\n\n## Minggu ini, buka *profiler* dulu\n\nJadi balik ke *endpoint* yang spike tadi. Sebelum nulis baris Rust pertama, buka *profiler* dulu minggu ini. Cari tahu apakah spike-nya datang dari GC, dari *query*, atau dari pool koneksi yang nggak dibagi. Kalau jawabannya satu dari dua yang pertama dan cuma bisa dibereskan dengan pindah *runtime*, kamu punya data buat meyakinkan tim. Kalau bukan, selamat, kamu baru aja hemat berbulan-bulan kerja. 😄\n","Pindah bahasa baru punya dasar kalau bottleneck-nya terbukti berasal dari runtime atau dari arsitektur yang nggak bisa berbagi resource, dan komponennya cukup terisolasi untuk digarap sendirian.\n- Yang sebenarnya kamu beli dari pindah bahasa bukan \"kode yang lebih cepat\", tapi hilangnya GC di kasus Discord dan kemampuan berbagi resource lintas thread di kasus Cloudflare.\n- Discord baru pindah ke Rust setelah dua usaha tuning-nya gagal, dan Cloudflare mengevaluasi pilihannya tiap kuartal selama beberapa tahun sebelum akhirnya membangun Pingora.\n- Instagram justru cukup mematikan GC di Python-nya dan langsung jadi 10% lebih efisien, jadi bahasa yang dianggap lambat pun bukan masalah selama dipakai dengan benar.\n- Selama masih ada query, indeks, atau konfigurasi yang belum dibereskan, masalah yang sama cuma pindah ke bahasa baru berikut biaya belajarnya.","\u003Cp>Bayangin ada satu \u003Cem>endpoint\u003C\u002Fem> yang grafik latency-nya nggak rata. Kebanyakan waktu dia adem, tapi tiba-tiba naik lancip, turun lagi, lalu naik lagi. Kamu mandangin grafik itu sambil ngopi, dan yang muncul di kepala \u003Cstrong>bukan profiler\u003C\u002Fstrong>. Yang muncul adalah kalimat, "kayaknya pindah bahasa aja deh." 😅\u003C\u002Fp>\n\u003Cp>Dorongan itu wajar banget, kok. Ada perusahaan besar yang beneran pindah bahasa dan hasilnya bagus. Tapi ada yang jarang kita perhatiin: dua cerita paling terkenal itu \u003Cstrong>nggak dimulai dari bahasa baru\u003C\u002Fstrong>. Mereka mulai dari angka, dari ngukur sampai kelihatan persis di mana yang mentok. Joel Spolsky pernah bilang bahwa \u003Ca href=\"https:\u002F\u002Fwww.joelonsoftware.com\u002F2000\u002F04\u002F06\u002Fthings-you-should-never-do-part-i\u002F\">menulis ulang dari nol adalah "the single worst strategic mistake that any software company can make"\u003C\u002Fa>, dan kita bakal lihat kenapa kalimat itu masih relevan.\u003C\u002Fp>\n\u003Cp>Yuk kita bedah tiga kasus nyata, plus satu kontra-contoh. Tujuannya bukan bilang pindah bahasa itu salah. Tujuannya biar kamu tahu kapan solusi murah udah habis, dan kapan baru boleh mikir \u003Cem>rewrite\u003C\u002Fem>.\u003C\u002Fp>\n\u003Ch2 id=\"kenapa-discord-akhirnya-pindah-dari-go-ke-rust\">Kenapa \u003Cem>Discord\u003C\u002Fem> akhirnya pindah dari Go ke Rust?\u003C\u002Fh2>\n\u003Cp>Yang mereka pindahkan namanya Read States. Tugasnya cuma nyatet channel dan pesan mana yang udah kamu baca. Sederhana, tapi layanan ini dipanggil tiap kali kita connect, tiap kali kirim pesan, dan tiap kali buka pesan. Jadi dia nempel di hampir semua hal yang kita lakuin di sana, lho.\u003C\u002Fp>\n\u003Cp>Skalanya gede, nih. Ada \u003Ca href=\"https:\u002F\u002Fdiscord.com\u002Fblog\u002Fwhy-discord-is-switching-from-go-to-rust\">miliaran Read State\u003C\u002Fa>, satu per user per channel. Tiap instance cache memegang jutaan user dan puluhan juta Read State. Per detiknya ada ratusan ribu update cache dan puluhan ribu write \u003Cem>database\u003C\u002Fem>.\u003C\u002Fp>\n\u003Cp>Gejalanya emang aneh tapi konsisten: spike latency dan CPU tiap sekitar 2 menit. Penyebabnya, Go memaksa \u003Cem>GC\u003C\u002Fem> jalan minimal tiap 2 menit, terlepas dari seberapa besar \u003Cem>heap\u003C\u002Fem>-nya tumbuh. Dari \u003Ca href=\"https:\u002F\u002Fdiscord.com\u002Fblog\u002Fwhy-discord-is-switching-from-go-to-rust\">grafik di blog mereka\u003C\u002Fa>, CPU naik dari sekitar 20% ke 30–40%. \u003Cem>Average response time\u003C\u002Fem> naik dari sekitar 0–2 ms ke sekitar 25 ms. Yang p95 melonjak dari sekitar 0–10 ms ke sekitar 300 ms. Catat ya, itu p95, bukan p99, dan semua angka ini hasil baca grafik, jadi sifatnya perkiraan.\u003C\u002Fp>\n\u003Cp>Nah, bagian pentingnya: mereka nggak langsung pindah. Mereka nyoba tuning dulu, dan dua-duanya gagal. Pertama, mengubah GC percent dari \u003Cem>endpoint\u003C\u002Fem>. Hasilnya nggak ngaruh, soalnya alokasi mereka nggak cukup cepat untuk memicu GC lebih sering. Kedua, memperkecil \u003Cem>LRU cache\u003C\u002Fem>. Spike-nya memang mengecil, tapi p99-nya naik karena cache jadi lebih sering \u003Cem>miss\u003C\u002Fem> dan bebannya pindah ke \u003Cem>database\u003C\u002Fem>.\u003C\u002Fp>\n\u003Cp>Jadi mereka udah ketemu tembok, tuh. Bukan tembok malas, tapi tembok yang diukur dan dicoba dobrak dua kali. Baru setelah itu mereka porting ke Rust, dan kata mereka latency-nya sama bagusnya dengan versi Go, dan \u003Cstrong>nggak ada spike sama sekali\u003C\u002Fstrong>. Dengan optimasi dasar aja versi Rust sudah mengalahkan versi Go yang di-tuning mati-matian. Setelah diprofil lagi, Rust menang di latency, CPU, dan memori.\u003C\u002Fp>\n\u003Cp>Beberapa hari kemudian kapasitas cache dinaikkan ke 8 juta Read State, hal yang nggak mungkin di versi Go karena GC-nya makin lama makin berat. Hasilnya, \u003Cem>average\u003C\u002Fem> response time jadi satuan mikrodetik, dan Max @mention jadi satuan milidetik. Tapi lihat catatan kaki mereka sendiri: mereka nggak berpikir kamu harus menulis ulang semuanya di Rust cuma karena bisa.\u003C\u002Fp>\n\u003Ch2 id=\"cloudflare-kodenya-lebih-cepat-atau-arsitekturnya\">Cloudflare: kodenya lebih cepat, atau arsitekturnya?\u003C\u002Fh2>\n\u003Cp>Cerita kedua datang dari Cloudflare, yang mengganti \u003Cem>NGINX\u003C\u002Fem> dengan \u003Cem>Pingora\u003C\u002Fem>. Menurut \u003Ca href=\"https:\u002F\u002Fblog.cloudflare.com\u002Fhow-we-built-pingora-the-proxy-that-connects-cloudflare-to-the-internet\u002F\">tulisan mereka soal Pingora\u003C\u002Fa>, sistem barunya melayani lebih dari 1 triliun \u003Cem>request\u003C\u002Fem> per hari. Dia cuma butuh sepertiga CPU dan memori dibanding infrastruktur proxy sebelumnya. Di produksi, hematnya 70% CPU dan 67% memori di beban yang sama. TTFB juga turun 5 ms di median dan 80 ms di p95.\u003C\u002Fp>\n\u003Cp>Sekarang bagian yang paling sering kelewat. Hemat itu \u003Cstrong>bukan karena kodenya lebih cepat\u003C\u002Fstrong>, lho. Layanan lama mereka juga udah sub-milidetik. Yang lebih cepat itu cara dia berbagi koneksi. 🤯\u003C\u002Fp>\n\u003Cp>Masalahnya ada di arsitektur NGINX, sih. Satu \u003Cem>request\u003C\u002Fem> cuma dilayani satu \u003Cem>worker\u003C\u002Fem>, yaitu satu proses OS. Load-nya nggak seimbang antar core, dan \u003Cem>connection pool\u003C\u002Fem>-nya per \u003Cem>worker\u003C\u002Fem>. Akibatnya makin banyak \u003Cem>worker\u003C\u002Fem> ditambah, makin jelek rasio \u003Cem>reuse\u003C\u002Fem> koneksinya. Ini kayak kalau tiap orang punya galon air sendiri di mejanya. Kamu butuh minum, galon sebelah masih penuh, tapi nggak bisa diambil, jadi kamu beli galon baru lagi. Tambah orang bukan bikin galon yang ada lebih kepakai, malah nambah galon baru terus.\u003C\u002Fp>\n\u003Cp>Pingora dibangun dengan multithreading, \u003Cem>runtime\u003C\u002Fem> async \u003Cem>Tokio\u003C\u002Fem>, dan \u003Cem>work stealing\u003C\u002Fem>. Mereka juga bikin HTTP \u003Cem>library\u003C\u002Fem> sendiri karena \u003Cem>traffic\u003C\u002Fem> di internet banyak yang nggak patuh RFC. Hasilnya, koneksi baru yang dibuat per detik tinggal sepertiga. Untuk satu pelanggan besar, rasio \u003Cem>reuse\u003C\u002Fem> koneksi naik dari 87,1% ke 99,92%, dan koneksi baru ke origin turun 160x. Terjemahan blog mereka: "kami menghemat 434 tahun waktu \u003Cem>handshake\u003C\u002Fem> setiap hari."\u003C\u002Fp>\n\u003Cp>Ada alasan lain mereka pindah. C nggak \u003Cem>memory safe\u003C\u002Fem>, Lua kurang performan dan nggak ada \u003Cem>static typing\u003C\u002Fem>, dan komunitas NGINX bergerak lambat. Setelah melayani beberapa ratus triliun \u003Cem>request\u003C\u002Fem>, Pingora belum pernah crash karena kode service-nya. Sistem ini lalu \u003Ca href=\"https:\u002F\u002Fblog.cloudflare.com\u002Fpingora-open-source\">di-open source-kan pada 28 Februari 2024\u003C\u002Fa>.\u003C\u002Fp>\n\u003Ch2 id=\"jadi-sebenarnya-apa-yang-kamu-beli-dengan-pindah-bahasa\">Jadi sebenarnya apa yang kamu beli dengan pindah bahasa?\u003C\u002Fh2>\n\u003Cp>Dua kemenangan tadi datang dari dua mekanisme yang beda. Di \u003Cem>Discord\u003C\u002Fem>, yang hilang adalah GC. Di Cloudflare, yang didapat adalah kemampuan berbagi resource lintas \u003Cem>thread\u003C\u002Fem>. Itu aja, sih. Bukan "Rust lebih cepat" secara magis, tapi ada \u003Cstrong>satu batasan spesifik\u003C\u002Fstrong> yang akhirnya lenyap.\u003C\u002Fp>\n\u003Cp>Kalau masalahmu bukan salah satu dari dua itu, ganti bahasa nggak akan mengubah apa pun. Query lambat tetap lambat di Rust. N+1 tetap N+1 di Go. Kamu cuma bakal punya masalah yang sama dengan sintaks yang berbeda, plus biaya belajar yang nggak kecil, lho.\u003C\u002Fp>\n\u003Cp>Kutipan Knuth sering dipotong jadi slogan. Versi lengkapnya begini: \u003Cem>"We should forget about small efficiencies, say about 97% of the time: premature optimization is the root of all evil. Yet we should not pass up our opportunities in that critical 3%."\u003C\u002Fem> Slogan itu sering dipakai buat menghindari ngukur, padahal Knuth justru minta kita tahu kode mana yang masuk 3% itu. Dan satu-satunya cara tahu adalah dengan \u003Cem>profiler\u003C\u002Fem>.\u003C\u002Fp>\n\u003Ch2 id=\"gimana-kalau-masalahmu-memang-gc-tapi-bahasanya-python\">Gimana kalau masalahmu memang GC, tapi bahasanya Python?\u003C\u002Fh2>\n\u003Cp>Di sini ada kabar yang menenangkan. Kalau masalahmu memang GC, kamu nggak selalu butuh bahasa baru. Pada Januari 2017, \u003Ca href=\"https:\u002F\u002Finstagram-engineering.com\u002Fdismissing-python-garbage-collection-at-instagram-4dca40b29172\">Instagram mematikan GC di Python-nya\u003C\u002Fa> dan langsung jadi 10% lebih efisien. Nggak pindah bahasa, nggak ganti \u003Cem>framework\u003C\u002Fem>, \u003Cstrong>cuma mematikan satu fitur\u003C\u002Fstrong>. 🙌\u003C\u002Fp>\n\u003Cp>Konteksnya gini. Instagram jalan dengan \u003Cem>Django\u003C\u002Fem> multi-proses, pakai \u003Cem>uWSGI\u003C\u002Fem> model \u003Cem>pre-fork\u003C\u002Fem>. Satu master mem-\u003Cem>fork\u003C\u002Fem> dirinya jadi puluhan \u003Cem>worker\u003C\u002Fem> yang melayani \u003Cem>request\u003C\u002Fem>. Idealnya, memori yang dibagi antar \u003Cem>worker\u003C\u002Fem> itu tetap dibagi.\u003C\u002Fp>\n\u003Cp>Tapi temuannya lain. \u003Cem>Shared memory\u003C\u002Fem> tiap \u003Cem>worker\u003C\u002Fem> turun dari sekitar 250 MB ke sekitar 140 MB, kira-kira sepertiga, cuma dalam beberapa detik setelah \u003Cem>worker\u003C\u002Fem> lahir. Sebabnya, Python pakai \u003Cem>reference counting\u003C\u002Fem>. Setiap kali sebuah objek dibaca, counter-nya ditulis. Tulisan itu memicu \u003Cem>copy-on-write\u003C\u002Fem> di kernel, jadi halaman memori yang tadinya dibagi berubah jadi milik masing-masing proses.\u003C\u002Fp>\n\u003Cp>Coba pikirin sebentar, deh. Gejala "GC bikin latency jelek" muncul di tiga tempat berbeda: di Go (\u003Cem>Discord\u003C\u002Fem>), di JVM Cassandra (\u003Cem>Discord\u003C\u002Fem> lagi), dan di Python (Instagram). Jawabannya beda-beda. \u003Cem>Discord\u003C\u002Fem> pindah bahasa, Instagram cuma ganti konfigurasi. Yang salah bukan bahasanya, tapi keputusan yang diambil di sekitar bahasanya.\u003C\u002Fp>\n\u003Ch2 id=\"kadang-bottleneck-nya-bahkan-bukan-bahasa\">Kadang bottleneck-nya bahkan bukan bahasa\u003C\u002Fh2>\n\u003Cp>Ini kontra-contohnya, dan lagi-lagi dari \u003Cem>Discord\u003C\u002Fem>. Pada 2023, mereka pindah dari \u003Cem>Cassandra\u003C\u002Fem> ke \u003Cem>ScyllaDB\u003C\u002Fem>, dan \u003Ca href=\"https:\u002F\u002Fdiscord.com\u002Fblog\u002Fhow-discord-stores-trillions-of-messages\">177 node jadi 72 node\u003C\u002Fa>. P99 untuk mengambil pesan historis turun dari 40–125 ms jadi 15 ms. Insert pesan turun dari 5–70 ms jadi 5 ms.\u003C\u002Fp>\n\u003Cp>Akar masalah aslinya adalah \u003Cem>hot partition\u003C\u002Fem>, konkurensi yang nggak terkendali, dan GC di JVM Cassandra. Yang menyelesaikannya adalah \u003Cem>request coalescing\u003C\u002Fem> plus \u003Cem>consistent hash routing\u003C\u002Fem> (ditulis di Rust), serta \u003Cem>database\u003C\u002Fem> yang bebas GC. Jadi bahkan di cerita "pindah ke Rust", sebagian besar kemenangannya datang dari desain dan pilihan komponen yang tepat.\u003C\u002Fp>\n\u003Cp>Bonusnya, migrator-nya ditulis di Rust dan estimasi migrasinya berubah dari 3 bulan jadi 9 hari. Puncaknya jalan di 3,2 juta pesan per detik. Jadi Rust tetap punya tempat yang jelas, tapi tempatnya sempit dan spesifik.\u003C\u002Fp>\n\u003Ch2 id=\"kapan-rewrite-boleh-dipertimbangkan\">Kapan \u003Cem>rewrite\u003C\u002Fem> boleh dipertimbangkan?\u003C\u002Fh2>\n\u003Cp>Alasan Spolsky masuk akal: kode lama itu sudah dipakai, sudah diuji, dan ratusan \u003Cem>bug\u003C\u002Fem> sudah ditemukan dan diperbaiki. Menulis ulang berarti membuang semua pengetahuan itu. Dia juga bilang soal performa, "1% of the work gets you 99% of the bang." Kalau 1% itu belum kamu cari, jangan loncat ke 100%.\u003C\u002Fp>\n\u003Cp>Jadi rewrite itu \u003Cstrong>keputusan terakhir\u003C\u002Fstrong>, bukan langkah pertama. Ada tiga pertanyaan yang harus kamu punya datanya.\u003C\u002Fp>\n\u003Cul>\n\u003Cli>Udah habis belum usaha yang lebih murah, seperti \u003Cem>query\u003C\u002Fem>, indeks, \u003Cem>caching\u003C\u002Fem>, N+1, dan \u003Cem>connection pool\u003C\u002Fem>? Ingat, Cloudflare mengevaluasi pilihannya tiap kuartal selama beberapa tahun.\u003C\u002Fli>\n\u003Cli>Bottleneck-nya kebukti berasal dari \u003Cem>runtime\u003C\u002Fem> atau arsitektur, bukan dari kode kita sendiri?\u003C\u002Fli>\n\u003Cli>Komponennya cukup terisolasi? \u003Cem>Discord\u003C\u002Fem> cuma memindahkan satu service, dan Cloudflare cuma mengganti layer proxy.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>Kalau belum bisa jawab ketiganya dengan data, jawabannya sekarang adalah belum. Dan itu jawaban yang sehat, kok, bukan tanda kamu pengecut.\u003C\u002Fp>\n\u003Ch2 id=\"kalau-tetap-harus-pindah-pilih-yang-mana\">Kalau tetap harus pindah, pilih yang mana?\u003C\u002Fh2>\n\u003Cp>Ini cuma pendapat, ya. Pilih Go kalau bebanmu didominasi I\u002FO dan kamu butuh cepat produktif. Pilih Rust kalau bebanmu CPU-bound, kamu memegang cache besar di RAM, kamu bangun infrastruktur jaringan atau storage, atau kamu butuh p99 yang datar tanpa toleransi jeda GC.\u003C\u002Fp>\n\u003Cp>Dan kalau Python atau Node-mu masih jalan baik setelah semua usaha murah dicoba, nggak ada yang memaksa kamu pindah. Bahasa yang dianggap lambat pun bukan masalah kalau dipakai dengan benar.\u003C\u002Fp>\n\u003Ch2 id=\"minggu-ini-buka-profiler-dulu\">Minggu ini, buka \u003Cem>profiler\u003C\u002Fem> dulu\u003C\u002Fh2>\n\u003Cp>Jadi balik ke \u003Cem>endpoint\u003C\u002Fem> yang spike tadi. Sebelum nulis baris Rust pertama, buka \u003Cem>profiler\u003C\u002Fem> dulu minggu ini. Cari tahu apakah spike-nya datang dari GC, dari \u003Cem>query\u003C\u002Fem>, atau dari pool koneksi yang nggak dibagi. Kalau jawabannya satu dari dua yang pertama dan cuma bisa dibereskan dengan pindah \u003Cem>runtime\u003C\u002Fem>, kamu punya data buat meyakinkan tim. Kalau bukan, selamat, kamu baru aja hemat berbulan-bulan kerja. 😄\u003C\u002Fp>\n","\u003Cp>Pindah bahasa baru punya dasar kalau bottleneck-nya terbukti berasal dari runtime atau dari arsitektur yang nggak bisa berbagi resource, dan komponennya cukup terisolasi untuk digarap sendirian.\u003C\u002Fp>\n\u003Cul>\n\u003Cli>Yang sebenarnya kamu beli dari pindah bahasa bukan "kode yang lebih cepat", tapi hilangnya GC di kasus Discord dan kemampuan berbagi resource lintas thread di kasus Cloudflare.\u003C\u002Fli>\n\u003Cli>Discord baru pindah ke Rust setelah dua usaha tuning-nya gagal, dan Cloudflare mengevaluasi pilihannya tiap kuartal selama beberapa tahun sebelum akhirnya membangun Pingora.\u003C\u002Fli>\n\u003Cli>Instagram justru cukup mematikan GC di Python-nya dan langsung jadi 10% lebih efisien, jadi bahasa yang dianggap lambat pun bukan masalah selama dipakai dengan benar.\u003C\u002Fli>\n\u003Cli>Selama masih ada query, indeks, atau konfigurasi yang belum dibereskan, masalah yang sama cuma pindah ke bahasa baru berikut biaya belajarnya.\u003C\u002Fli>\n\u003C\u002Ful>\n",[708,725,738,751,764,784],{"id":709,"title":710,"title_en":21,"slug":711,"excerpt":712,"excerpt_en":21,"has_translation":15,"published_at":713,"reading_time":632,"hidden_from_listing":15,"tags":714,"thumbnail_image":21,"og_image":716,"series":717,"authors":722},133,"Python Strings: the Badge That Only Worked for One Name","python-strings","The badge script printed the right name for the person who wrote it and the wrong name for everybody else. Slicing counts characters, and a name is words.","2026-10-28T13:00:00.000Z",[260,715],"string","\u002Fog\u002Fpython-strings.png",{"id":718,"title":719,"title_en":720,"slug":721},30,"Koleksi dan String","Collections and Strings","python-collections",[723],{"authors_id":724},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"id":726,"title":727,"title_en":21,"slug":728,"excerpt":729,"excerpt_en":21,"has_translation":15,"published_at":730,"reading_time":632,"hidden_from_listing":15,"tags":731,"thumbnail_image":21,"og_image":733,"series":734,"authors":735},132,"Python Tuples: Why All Three Points Are the Same Point","python-tuples","Three points walk into a loop and come out identical. The list was never copied three times; the name was. A tuple makes that bug impossible.","2026-10-28T12:00:00.000Z",[260,732],"tuple","\u002Fog\u002Fpython-tuples.png",{"id":718,"title":719,"title_en":720,"slug":721},[736],{"authors_id":737},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"id":739,"title":740,"title_en":21,"slug":741,"excerpt":742,"excerpt_en":21,"has_translation":15,"published_at":743,"reading_time":632,"hidden_from_listing":15,"tags":744,"thumbnail_image":21,"og_image":746,"series":747,"authors":748},131,"Python Sets: the Duplicate That Keeps Coming Back","python-sets","A de-duplication loop that works on clean data leaves four tags where there should be two. The bug is case, and the fix is a `set`.","2026-10-28T11:00:00.000Z",[260,745],"set","\u002Fog\u002Fpython-sets.png",{"id":718,"title":719,"title_en":720,"slug":721},[749],{"authors_id":750},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"id":752,"title":753,"title_en":21,"slug":754,"excerpt":755,"excerpt_en":21,"has_translation":15,"published_at":756,"reading_time":632,"hidden_from_listing":15,"tags":757,"thumbnail_image":21,"og_image":759,"series":760,"authors":761},130,"Python Dictionaries: the Key That Isn't There","python-dictionary","A word counter dies the first time it meets a word twice. The fix is one line; what takes longer is knowing which of a dict's three lookups to reach for.","2026-10-28T10:00:00.000Z",[260,758],"dictionary","\u002Fog\u002Fpython-dictionary.png",{"id":718,"title":719,"title_en":720,"slug":721},[762],{"authors_id":763},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"id":765,"title":766,"title_en":767,"slug":768,"excerpt":769,"excerpt_en":770,"has_translation":12,"published_at":771,"reading_time":14,"hidden_from_listing":15,"tags":772,"thumbnail_image":21,"og_image":779,"series":780,"authors":781},129,"Biar RAG nggak kena 504: pisahin jalur query dan ingestion","So Your RAG Stops Hitting 504: Split the Query Path From Ingestion","rag-pisah-jalur-query-dan-ingestion","Satu service RAG yang merangkap parsing dan menjawab kueri bakal tumbang pas ada yang upload PDF 200 halaman. Obatnya dua jalur dan satu queue.","One RAG service that both parses uploads and answers queries will fall over the moment someone uploads a 200-page PDF. The cure is two paths and one queue.","2026-10-22T01:30:00.000Z",[773,774,775,776,777,778],"rag","cloud-run","cloud-tasks","gcp","serverless","arsitektur","\u002Fog\u002Frag-pisah-jalur-query-dan-ingestion.png",{"id":694,"title":35,"title_en":36,"slug":37},[782],{"authors_id":783},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},{"id":785,"title":786,"title_en":787,"slug":788,"excerpt":789,"excerpt_en":790,"has_translation":12,"published_at":791,"reading_time":632,"hidden_from_listing":15,"tags":792,"thumbnail_image":21,"og_image":799,"series":800,"authors":801},128,"Linear atau graf: cara mengatur banyak agent tanpa loop abadi","Linear or Graph: How to Run Many Agents Without an Eternal Loop","multi-agen-linear-atau-grafik","Satu agent dengan dua puluh tool gampang kacau. Begitu beberapa agent saling revisi, kamu butuh graf yang mendukung cycle dan angka yang membuatnya berhenti.","One agent with twenty tools gets messy fast. Once several agents revise each other, you need a graph that allows cycles and a number that makes it stop.","2026-10-21T16:05:00.000Z",[793,794,795,796,797,798],"multi-agent","langgraph","autogen","orkestrasi","checkpoint","llm","\u002Fog\u002Fmulti-agen-linear-atau-grafik.png",{"id":694,"title":35,"title_en":36,"slug":37},[802],{"authors_id":803},{"id":306,"name":307,"slug":308,"role":271,"profile_picture":21},[805,808,811,814,817,820,823,826],{"text":806,"depth":268,"id":807},"Kenapa Discord akhirnya pindah dari Go ke Rust?","kenapa-discord-akhirnya-pindah-dari-go-ke-rust",{"text":809,"depth":268,"id":810},"Cloudflare: kodenya lebih cepat, atau arsitekturnya?","cloudflare-kodenya-lebih-cepat-atau-arsitekturnya",{"text":812,"depth":268,"id":813},"Jadi sebenarnya apa yang kamu beli dengan pindah bahasa?","jadi-sebenarnya-apa-yang-kamu-beli-dengan-pindah-bahasa",{"text":815,"depth":268,"id":816},"Gimana kalau masalahmu memang GC, tapi bahasanya Python?","gimana-kalau-masalahmu-memang-gc-tapi-bahasanya-python",{"text":818,"depth":268,"id":819},"Kadang bottleneck-nya bahkan bukan bahasa","kadang-bottleneck-nya-bahkan-bukan-bahasa",{"text":821,"depth":268,"id":822},"Kapan rewrite boleh dipertimbangkan?","kapan-rewrite-boleh-dipertimbangkan",{"text":824,"depth":268,"id":825},"Kalau tetap harus pindah, pilih yang mana?","kalau-tetap-harus-pindah-pilih-yang-mana",{"text":827,"depth":268,"id":828},"Minggu ini, buka profiler dulu","minggu-ini-buka-profiler-dulu",1793295833851]