Apache Airflow: Veri Ak\u0131\u015flar\u0131n\u0131 Otomatikle\u015ftirmek \u0130\u00e7in Kapsaml\u0131 Rehber<\/title> \n<body><\/p>\n<style>\n \/* Temel stil tan\u0131mlamalar\u0131 *\/\n body {\n font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;\n line-height: 1.6;\n color: #333;\n margin: 0 auto;\n padding: 20px;\n max-width: 1200px;\n background-color: #f9f9f9;\n }\n h2, h3 {\n color: #0056b3;\n margin-top: 30px;\n margin-bottom: 15px;\n }\n p {\n margin-bottom: 1em;\n }\n a {\n color: #007bff;\n text-decoration: none;\n }\n a:hover {\n text-decoration: underline;\n }\n ul, ol {\n margin-bottom: 1em;\n padding-left: 20px;\n }\n li {\n margin-bottom: 0.5em;\n }\n pre {\n background-color: #eee;\n border: 1px solid #ddd;\n padding: 15px;\n border-radius: 5px;\n overflow-x: auto;\n margin-bottom: 1.5em;\n }\n code {\n font-family: 'Consolas', 'Monaco', monospace;\n font-size: 0.9em;\n }\n table {\n width: 100%;\n border-collapse: collapse;\n margin-bottom: 1.5em;\n }\n th, td {\n border: 1px solid #ddd;\n padding: 8px;\n text-align: left;\n }\n th {\n background-color: #f2f2f2;\n font-weight: bold;\n }\n .uzman-ipucu, .onemli-not {\n background-color: #e6f7ff;\n border-left: 5px solid #007bff;\n padding: 15px;\n margin: 20px 0;\n border-radius: 4px;\n font-style: italic;\n }\n .onemli-not {\n background-color: #fff3cd;\n border-left: 5px solid #ffc107;\n color: #856404;\n }<\/p>\n \/* Mobil uyumluluk i\u00e7in medya sorgular\u0131 *\/\n @media (max-width: 768px) {\n body {\n padding: 15px;\n }\n h2 {\n font-size: 1.8em;\n }\n h3 {\n font-size: 1.4em;\n }\n table, thead, tbody, th, td, tr {\n display: block;\n }\n thead tr {\n position: absolute;\n top: -9999px;\n left: -9999px;\n }\n tr {\n border: 1px solid #ccc;\n margin-bottom: 10px;\n }\n td {\n border: none;\n border-bottom: 1px solid #eee;\n position: relative;\n padding-left: 50%;\n text-align: right;\n }\n td:before {\n position: absolute;\n top: 6px;\n left: 6px;\n width: 45%;\n padding-right: 10px;\n white-space: nowrap;\n text-align: left;\n font-weight: bold;\n }\n \/* Her s\u00fctun i\u00e7in ba\u015fl\u0131k etiketi *\/\n td:nth-of-type(1):before { content: \"Bile\u015fen\"; }\n td:nth-of-type(2):before { content: \"G\u00f6revi\"; }\n }<\/p>\n @media (max-width: 480px) {\n body {\n padding: 10px;\n }\n pre {\n padding: 10px;\n font-size: 0.8em;\n }\n }\n<\/style>\nVeri ak\u0131\u015flar\u0131n\u0131z\u0131n karma\u015f\u0131kl\u0131\u011f\u0131yla bo\u011fu\u015fuyor, manuel m\u00fcdahalelerle zaman kaybediyor veya hata takibinde zorlan\u0131yor musunuz? Apache Airflow, veri m\u00fchendisli\u011fi s\u00fcre\u00e7lerinizi, ETL i\u015f ak\u0131\u015flar\u0131n\u0131z\u0131 ve di\u011fer zamanlanm\u0131\u015f g\u00f6revlerinizi Python koduyla programatik olarak y\u00f6netmenizi sa\u011flayan a\u00e7\u0131k kaynakl\u0131 bir platformdur. Bu rehber, Airflow’un temel prensiplerinden ba\u015flayarak, nas\u0131l \u00e7al\u0131\u015ft\u0131\u011f\u0131n\u0131, neden bu kadar pop\u00fcler oldu\u011funu ve kendi veri orkestrasyon \u00e7\u00f6z\u00fcmlerinizi nas\u0131l olu\u015fturaca\u011f\u0131n\u0131z\u0131 ad\u0131m ad\u0131m a\u00e7\u0131klayacak.<\/p>\n<h2>Apache Airflow Nedir ve Neden Veri Ak\u0131\u015flar\u0131n\u0131z \u0130\u00e7in Kritik Bir Ara\u00e7t\u0131r?<\/h2>\nModern veri odakl\u0131 d\u00fcnyada, \u015firketler s\u00fcrekli artan miktarda veriyi i\u015flemek, analiz etmek ve anlaml\u0131 i\u00e7g\u00f6r\u00fclere d\u00f6n\u00fc\u015ft\u00fcrmek zorundad\u0131r. Bu s\u00fcre\u00e7 genellikle farkl\u0131 kaynaklardan veri \u00e7ekme (Extract), d\u00f6n\u00fc\u015ft\u00fcrme (Transform) ve bir hedefe y\u00fckleme (Load) ad\u0131mlar\u0131n\u0131 i\u00e7eren karma\u015f\u0131k ETL (Extract, Transform, Load) boru hatlar\u0131n\u0131 i\u00e7erir. Ancak bu boru hatlar\u0131, birden fazla ba\u011f\u0131ml\u0131 g\u00f6revin, farkl\u0131 sistemlerle etkile\u015fimlerin ve hata y\u00f6netiminin bir araya gelmesiyle olduk\u00e7a karma\u015f\u0131k hale gelebilir. \u0130\u015fte tam da bu noktada Apache Airflow devreye giriyor.<\/p>\nApache Airflow, programatik olarak i\u015f ak\u0131\u015flar\u0131n\u0131 yazmak, zamanlamak ve izlemek i\u00e7in tasarlanm\u0131\u015f a\u00e7\u0131k kaynakl\u0131 bir platformdur. Basit\u00e7e ifade etmek gerekirse, bir dizi g\u00f6revi belirli bir s\u0131rada ve belirli ko\u015fullar alt\u0131nda \u00e7al\u0131\u015ft\u0131rman\u0131za olanak tan\u0131yan bir “i\u015f ak\u0131\u015f\u0131 orkestrasyon” arac\u0131d\u0131r. Bu i\u015f ak\u0131\u015flar\u0131, y\u00f6nlendirilmi\u015f d\u00f6ng\u00fcsel olmayan grafikler (Directed Acyclic Graphs – DAGs) olarak tan\u0131mlan\u0131r ve tamamen Python koduyla yaz\u0131l\u0131r. Bu, Airflow’u son derece esnek, \u00f6l\u00e7eklenebilir ve geli\u015ftiriciler i\u00e7in eri\u015filebilir k\u0131lar. Geleneksel cron tabanl\u0131 zamanlama ara\u00e7lar\u0131n\u0131n aksine, Airflow g\u00f6rev ba\u011f\u0131ml\u0131l\u0131klar\u0131n\u0131, yeniden denemeleri, hata bildirimlerini ve izlemeyi yerle\u015fik olarak sunar, bu da veri m\u00fchendislerinin hayat\u0131n\u0131 \u00f6nemli \u00f6l\u00e7\u00fcde kolayla\u015ft\u0131r\u0131r.<\/p>\nPeki, Airflow neden bu kadar kritik bir ara\u00e7 haline geldi? \u00d6ncelikle, veri hacminin ve \u00e7e\u015fitlili\u011finin artmas\u0131yla birlikte, manuel s\u00fcre\u00e7ler s\u00fcrd\u00fcr\u00fclemez hale gelmi\u015ftir. Airflow, bu s\u00fcre\u00e7leri otomatikle\u015ftirerek insan hatas\u0131n\u0131 minimize eder ve operasyonel verimlili\u011fi art\u0131r\u0131r. \u0130kincisi, veri ak\u0131\u015flar\u0131n\u0131n karma\u015f\u0131kl\u0131\u011f\u0131, g\u00f6revler aras\u0131ndaki ba\u011f\u0131ml\u0131l\u0131klar\u0131n do\u011fru bir \u015fekilde y\u00f6netilmesini gerektirir. Airflow’un DAG yap\u0131s\u0131, bu ba\u011f\u0131ml\u0131l\u0131klar\u0131 a\u00e7\u0131k\u00e7a tan\u0131mlaman\u0131za ve garantilemenize olanak tan\u0131r. \u00d6rne\u011fin, bir veri d\u00f6n\u00fc\u015f\u00fcm g\u00f6revinin, ancak ilgili veri \u00e7ekme g\u00f6revi ba\u015far\u0131yla tamamland\u0131ktan sonra ba\u015flamas\u0131n\u0131 sa\u011flayabilirsiniz. \u00dc\u00e7\u00fcnc\u00fcs\u00fc, hata y\u00f6netimi ve izleme, b\u00fcy\u00fck \u00f6l\u00e7ekli veri boru hatlar\u0131nda hayati \u00f6neme sahiptir. Airflow’un zengin kullan\u0131c\u0131 aray\u00fcz\u00fc (UI), t\u00fcm i\u015f ak\u0131\u015flar\u0131n\u0131z\u0131n durumunu ger\u00e7ek zamanl\u0131 olarak g\u00f6rmenizi, ba\u015far\u0131s\u0131z olan g\u00f6revleri kolayca tespit etmenizi ve hatta manuel olarak yeniden \u00e7al\u0131\u015ft\u0131rman\u0131z\u0131 sa\u011flar. Bu sayede, sorunlara h\u0131zl\u0131ca m\u00fcdahale edebilir ve veri tutarl\u0131l\u0131\u011f\u0131n\u0131 koruyabilirsiniz. Son olarak, Python tabanl\u0131 olmas\u0131, veri bilimcileri ve m\u00fchendislerinin zaten a\u015fina oldu\u011fu bir dilde i\u015f ak\u0131\u015flar\u0131n\u0131 tan\u0131mlamas\u0131na olanak tan\u0131r, bu da \u00f6\u011frenme e\u011frisini d\u00fc\u015f\u00fcr\u00fcr ve entegrasyonu kolayla\u015ft\u0131r\u0131r. T\u00fcm bu nedenler, Apache Airflow’u modern veri platformlar\u0131n\u0131n vazge\u00e7ilmez bir par\u00e7as\u0131 haline getiriyor.<\/p>\n<h3>Apache Airflow’un Temel Bile\u015fenleri Nelerdir ve Nas\u0131l \u00c7al\u0131\u015f\u0131rlar?<\/h3>\nApache Airflow’un arkas\u0131ndaki g\u00fcc\u00fc anlamak i\u00e7in, onun temel bile\u015fenlerini ve bu bile\u015fenlerin birbiriyle nas\u0131l etkile\u015fim kurdu\u011funu bilmek \u00f6nemlidir. Airflow, da\u011f\u0131t\u0131k bir sistem olarak \u00e7al\u0131\u015fabilir ve genellikle a\u015fa\u011f\u0131daki ana bile\u015fenlerden olu\u015fur:<\/p>\n<ol>\n<li>Webserver (Web Sunucusu):<\/strong> Airflow’un kullan\u0131c\u0131 aray\u00fcz\u00fcn\u00fc (UI) bar\u0131nd\u0131ran bile\u015fendir. Bu aray\u00fcz sayesinde DAG’lar\u0131n\u0131z\u0131 g\u00f6rselle\u015ftirebilir, g\u00f6revlerin durumunu izleyebilir, ge\u00e7mi\u015f \u00e7al\u0131\u015ft\u0131rmalar\u0131 g\u00f6r\u00fcnt\u00fcleyebilir, manuel olarak tetikleyebilir ve genel Airflow ortam\u0131n\u0131z\u0131 y\u00f6netebilirsiniz. Bu, veri m\u00fchendisleri ve operasyon ekipleri i\u00e7in i\u015f ak\u0131\u015flar\u0131n\u0131n sa\u011fl\u0131\u011f\u0131n\u0131 ve performans\u0131n\u0131 anlamak ad\u0131na merkezi bir noktad\u0131r.<\/li>\n<li>Scheduler (Zamanlay\u0131c\u0131):<\/strong> Airflow’un kalbidir. DAG’lar\u0131 d\u00fczenli aral\u0131klarla tarar, zamanlanm\u0131\u015f g\u00f6rev \u00f6rneklerini (Task Instances) olu\u015fturur ve bu g\u00f6revlerin \u00e7al\u0131\u015ft\u0131r\u0131lmas\u0131 i\u00e7in uygun zamanda Worker’lara g\u00f6nderir. Scheduler, g\u00f6rev ba\u011f\u0131ml\u0131l\u0131klar\u0131n\u0131, yeniden deneme politikalar\u0131n\u0131 ve zaman pencerelerini y\u00f6netir. S\u00fcrekli olarak \u00e7al\u0131\u015f\u0131r ve Airflow’un t\u00fcm i\u015f ak\u0131\u015flar\u0131n\u0131 canl\u0131 tutar.<\/li>\n<li>Worker (\u0130\u015f\u00e7i):<\/strong> Scheduler taraf\u0131ndan g\u00f6nderilen g\u00f6revleri (Task Instances) fiilen \u00e7al\u0131\u015ft\u0131ran bile\u015fenlerdir. Airflow, farkl\u0131 y\u00fcr\u00fct\u00fcc\u00fc (Executor) t\u00fcrlerini destekler (SequentialExecutor, LocalExecutor, CeleryExecutor, KubernetesExecutor vb.). \u00d6zellikle CeleryExecutor veya KubernetesExecutor gibi da\u011f\u0131t\u0131k y\u00fcr\u00fct\u00fcc\u00fcler kullan\u0131ld\u0131\u011f\u0131nda, birden fazla Worker paralel olarak \u00e7al\u0131\u015fabilir, bu da Airflow’un \u00f6l\u00e7eklenebilirli\u011fini art\u0131r\u0131r. Her Worker, kendisine atanan g\u00f6revi ba\u011f\u0131ms\u0131z olarak y\u00fcr\u00fct\u00fcr ve sonu\u00e7lar\u0131 veritaban\u0131na kaydeder.<\/li>\n<li>Database (Veritaban\u0131):<\/strong> Airflow’un t\u00fcm meta verilerini depolad\u0131\u011f\u0131 yerdir. Bu veriler aras\u0131nda DAG’lar\u0131n yap\u0131lar\u0131, g\u00f6revlerin durumlar\u0131, \u00e7al\u0131\u015ft\u0131rma ge\u00e7mi\u015fleri, ba\u011flant\u0131 bilgileri, de\u011fi\u015fkenler ve XCom (g\u00f6revler aras\u0131 ileti\u015fim) verileri bulunur. PostgreSQL veya MySQL gibi ili\u015fkisel veritabanlar\u0131 genellikle kullan\u0131l\u0131r. Veritaban\u0131, t\u00fcm Airflow bile\u015fenleri aras\u0131nda tek tutarl\u0131l\u0131k kayna\u011f\u0131d\u0131r ve sistemin do\u011fru \u00e7al\u0131\u015fmas\u0131 i\u00e7in hayati \u00f6neme sahiptir.<\/li>\n<\/ol>\nBu bile\u015fenler aras\u0131ndaki etkile\u015fim, Airflow’un sorunsuz \u00e7al\u0131\u015fmas\u0131n\u0131 sa\u011flar. Scheduler, DAG dosyalar\u0131n\u0131 tarar ve veritaban\u0131na kaydeder. Zaman\u0131 geldi\u011finde, veritaban\u0131ndaki bilgilere dayanarak g\u00f6rev \u00f6rneklerini olu\u015fturur ve bunlar\u0131 bir Worker’a atar. Worker, g\u00f6revi y\u00fcr\u00fct\u00fcrken durum g\u00fcncellemelerini ve \u00e7\u0131kt\u0131lar\u0131 veritaban\u0131na geri yazar. Webserver ise bu veritaban\u0131 bilgilerini \u00e7ekerek kullan\u0131c\u0131 aray\u00fcz\u00fcnde g\u00f6sterir. Bu mod\u00fcler yap\u0131, Airflow’un farkl\u0131 da\u011f\u0131t\u0131m senaryolar\u0131na (tek bir sunucuda veya da\u011f\u0131t\u0131k bir k\u00fcmede) kolayca adapte olmas\u0131n\u0131 ve farkl\u0131 i\u015f y\u00fckleri i\u00e7in \u00f6l\u00e7eklenmesini m\u00fcmk\u00fcn k\u0131lar. \u00d6rne\u011fin, k\u00fc\u00e7\u00fck bir proje i\u00e7in LocalExecutor ile tek bir sunucuda t\u00fcm bile\u015fenleri \u00e7al\u0131\u015ft\u0131rabilirken, b\u00fcy\u00fck \u00f6l\u00e7ekli ve y\u00fcksek performans gerektiren ortamlar i\u00e7in Celery veya Kubernetes tabanl\u0131 da\u011f\u0131t\u0131k bir mimari tercih edilebilir.<\/p>\n<div class=\"onemli-not\">\n \u00d6nemli Not: Airflow’un g\u00fcc\u00fc, bu bile\u015fenlerin uyumlu \u00e7al\u0131\u015fmas\u0131ndan gelir. Bir bile\u015fenin ar\u0131zalanmas\u0131 t\u00fcm sistemi etkileyebilir, bu y\u00fczden \u00f6zellikle \u00fcretim ortamlar\u0131nda her bir bile\u015fenin izlenmesi ve y\u00fcksek eri\u015filebilirli\u011finin sa\u011flanmas\u0131 kritik \u00f6neme sahiptir.\n<\/div>\n<h2>\u0130lk DAG’\u0131n\u0131z\u0131 Olu\u015fturmak: Ad\u0131m Ad\u0131m Rehber<\/h2>\nApache Airflow’un kalbi olan DAG’lar\u0131 (Directed Acyclic Graphs) anlamak ve olu\u015fturmak, platformu kullanmaya ba\u015flaman\u0131n ilk ve en \u00f6nemli ad\u0131m\u0131d\u0131r. DAG’lar, \u00e7al\u0131\u015ft\u0131rmak istedi\u011finiz g\u00f6revlerin ve bu g\u00f6revler aras\u0131ndaki ba\u011f\u0131ml\u0131l\u0131klar\u0131n tan\u0131mland\u0131\u011f\u0131 Python dosyalar\u0131d\u0131r. “Y\u00f6nlendirilmi\u015f” (Directed) olmas\u0131, g\u00f6revlerin belirli bir s\u0131raya g\u00f6re akaca\u011f\u0131n\u0131, “D\u00f6ng\u00fcsel Olmayan” (Acyclic) olmas\u0131 ise bir g\u00f6revin kendisine veya daha \u00f6nce \u00e7al\u0131\u015fm\u0131\u015f bir g\u00f6reve geri d\u00f6nemeyece\u011fi anlam\u0131na gelir, bu da sonsuz d\u00f6ng\u00fcleri engeller. Bu b\u00f6l\u00fcmde, basit bir “Merhaba D\u00fcnya” DAG’\u0131 olu\u015fturarak Airflow ile tan\u0131\u015faca\u011f\u0131z.<\/p>\n\u00d6ncelikle, Airflow ortam\u0131n\u0131z\u0131n kurulu ve \u00e7al\u0131\u015f\u0131r durumda oldu\u011funu varsay\u0131yoruz. Kurulum hakk\u0131nda detayl\u0131 bilgi i\u00e7in resmi Airflow dok\u00fcmantasyonuna ba\u015fvurabilirsiniz. Genellikle, bir Airflow kurulumu, bir <code class=\"language-\">airflow.cfg<\/code> dosyas\u0131, bir <code class=\"language-\">dags<\/code> klas\u00f6r\u00fc ve bir veritaban\u0131 ba\u011flant\u0131s\u0131 i\u00e7erir. DAG dosyalar\u0131n\u0131z\u0131 bu <code class=\"language-\">dags<\/code> klas\u00f6r\u00fcn\u00fcn i\u00e7ine yerle\u015ftirmeniz gerekmektedir.<\/p>\n\u015eimdi basit bir DAG olu\u015ftural\u0131m. Bu DAG, iki Python g\u00f6revi ve bir Bash g\u00f6revi i\u00e7erecek ve belirli bir s\u0131rayla \u00e7al\u0131\u015facakt\u0131r.<\/p>\n<pre><code>\nfrom airflow import DAG\nfrom airflow.operators.bash import BashOperator\nfrom airflow.operators.python import PythonOperator\nfrom datetime import datetime, timedelta\n\n# Varsay\u0131lan arg\u00fcmanlar, DAG'daki t\u00fcm g\u00f6revler i\u00e7in ge\u00e7erli olabilir\ndefault_args = {\n 'owner': 'airflow',\n 'depends_on_past': False,\n 'email': ['airflow@example.com'],\n 'email_on_failure': False,\n 'email_on_retry': False,\n 'retries': 1,\n 'retry_delay': timedelta(minutes=5),\n}\n\n# Python fonksiyonlar\u0131 tan\u0131mlayal\u0131m\ndef print_hello():\n print(\"Merhaba Airflow!\")\n return \"Hello\"\n\ndef print_world(ti):\n # XCom kullanarak \u00f6nceki g\u00f6revden veri \u00e7ekme\n message = ti.xcom_pull(task_ids='hello_task')\n print(f\"D\u00fcnya! \u00d6nceki g\u00f6revden gelen mesaj: {message}\")\n\n# DAG'\u0131 tan\u0131mlayal\u0131m\nwith DAG(\n dag_id='ilk_airflow_dag',\n default_args=default_args,\n description='Bu ilk Airflow DAG\\'\u0131m\u0131zd\u0131r.',\n schedule_interval=timedelta(days=1), # G\u00fcnde bir kez \u00e7al\u0131\u015ft\u0131r\n start_date=datetime(2023, 1, 1),\n catchup=False, # Ge\u00e7mi\u015fteki \u00e7al\u0131\u015ft\u0131rmalar\u0131 otomatik olarak tetikleme\n tags=['ornek', 'baslangic'],\n) as dag:\n # G\u00f6rev 1: Bash komutu \u00e7al\u0131\u015ft\u0131rma\n start_task = BashOperator(\n task_id='bash_baslangic_gorevi',\n bash_command='echo \"Airflow ile bash g\u00f6revi ba\u015flad\u0131!\"',\n )\n\n # G\u00f6rev 2: Python fonksiyonu \u00e7al\u0131\u015ft\u0131rma\n hello_task = PythonOperator(\n task_id='hello_task',\n python_callable=print_hello,\n )\n\n # G\u00f6rev 3: Ba\u015fka bir Python fonksiyonu \u00e7al\u0131\u015ft\u0131rma\n world_task = PythonOperator(\n task_id='world_task',\n python_callable=print_world,\n )\n\n # G\u00f6rev 4: Bash komutu \u00e7al\u0131\u015ft\u0131rma\n end_task = BashOperator(\n task_id='bash_bitis_gorevi',\n bash_command='echo \"Airflow ile bash g\u00f6revi tamamland\u0131!\"',\n )\n\n # G\u00f6rev ba\u011f\u0131ml\u0131l\u0131klar\u0131n\u0131 tan\u0131mlama\n # start_task -> hello_task -> world_task -> end_task\n start_task >> hello_task\n hello_task >> world_task\n world_task >> end_task\n<\/code><\/pre>\nBu kod blo\u011funu <code class=\"language-\">ilk_airflow_dag.py<\/code> ad\u0131yla Airflow’un <code class=\"language-\">dags<\/code> klas\u00f6r\u00fcne kaydetti\u011finizde, Scheduler otomatik olarak onu alg\u0131layacak ve Airflow UI’da g\u00f6r\u00fcn\u00fcr hale getirecektir. \u015eimdi kodu ve yap\u0131s\u0131n\u0131 biraz daha detayland\u0131ral\u0131m:<\/p>\n<ul>\n<li><code class=\"language-\">from airflow import DAG<\/code> ve Di\u011fer \u0130\u00e7e Aktarmalar:<\/strong> Airflow DAG’lar\u0131 ve operat\u00f6rleri i\u00e7in gerekli s\u0131n\u0131flar\u0131 i\u00e7e aktar\u0131yoruz. <code class=\"language-\">BashOperator<\/code> bir Bash komutu \u00e7al\u0131\u015ft\u0131rmak i\u00e7in, <code class=\"language-\">PythonOperator<\/code> ise bir Python fonksiyonu \u00e7al\u0131\u015ft\u0131rmak i\u00e7in kullan\u0131l\u0131r.<\/li>\n<li><code class=\"language-\">default_args<\/code>:<\/strong> Bu s\u00f6zl\u00fck, DAG i\u00e7indeki t\u00fcm g\u00f6revlere uygulanacak varsay\u0131lan parametreleri i\u00e7erir. \u00d6rne\u011fin, <code class=\"language-\">owner<\/code> (sahip), <code class=\"language-\">retries<\/code> (yeniden deneme say\u0131s\u0131) ve <code class=\"language-\">retry_delay<\/code> (yeniden deneme gecikmesi) gibi ayarlar burada tan\u0131mlanabilir. Bu, her g\u00f6reve tek tek ayn\u0131 ayarlar\u0131 yazma ihtiyac\u0131n\u0131 ortadan kald\u0131r\u0131r.<\/li>\n<li><code class=\"language-\">print_hello()<\/code> ve <code class=\"language-\">print_world(ti)<\/code> Fonksiyonlar\u0131:<\/strong> <code class=\"language-\">PythonOperator<\/code> taraf\u0131ndan \u00e7a\u011fr\u0131lacak basit Python fonksiyonlar\u0131d\u0131r. <code class=\"language-\">print_world<\/code> fonksiyonu, <code class=\"language-\">ti<\/code> (task instance) parametresini alarak, Airflow’un g\u00f6revler aras\u0131 ileti\u015fim mekanizmas\u0131 olan XCom’u kullanarak <code class=\"language-\">hello_task<\/code> g\u00f6revinden d\u00f6nen de\u011feri \u00e7ekiyor. Bu, karma\u015f\u0131k i\u015f ak\u0131\u015flar\u0131nda g\u00f6revler aras\u0131nda veri payla\u015f\u0131m\u0131 i\u00e7in \u00e7ok \u00f6nemlidir.<\/li>\n<li><code class=\"language-\">with DAG(...) as dag:<\/code>:<\/strong> Bu ba\u011flam y\u00f6neticisi (context manager) i\u00e7inde DAG’\u0131m\u0131z\u0131n tan\u0131m\u0131n\u0131 yap\u0131yoruz.\n<ul>\n<li><code class=\"language-\">dag_id<\/code>: DAG’\u0131n benzersiz tan\u0131mlay\u0131c\u0131s\u0131d\u0131r.<\/li>\n<li><code class=\"language-\">description<\/code>: DAG hakk\u0131nda k\u0131sa bir a\u00e7\u0131klama.<\/li>\n<li><code class=\"language-\">schedule_interval<\/code>: DAG’\u0131n ne s\u0131kl\u0131kla \u00e7al\u0131\u015ft\u0131r\u0131laca\u011f\u0131n\u0131 belirler. <code class=\"language-\">timedelta(days=1)<\/code> g\u00fcnde bir kez anlam\u0131na gelir. Cron ifadeleri (<code class=\"language-\">'0 0 * * *'<\/code>) de kullan\u0131labilir.<\/li>\n<li><code class=\"language-\">start_date<\/code>: DAG’\u0131n ilk ne zaman \u00e7al\u0131\u015fmaya ba\u015flayaca\u011f\u0131n\u0131 belirler.<\/li>\n<li><code class=\"language-\">catchup=False<\/code>: Bu \u00f6nemli bir parametredir. E\u011fer <code class=\"language-\">start_date<\/code> ge\u00e7mi\u015fte bir tarihte ve <code class=\"language-\">catchup=True<\/code> olsayd\u0131, Airflow, <code class=\"language-\">start_date<\/code> ile mevcut tarih aras\u0131ndaki t\u00fcm ka\u00e7\u0131r\u0131lm\u0131\u015f \u00e7al\u0131\u015ft\u0131rmalar\u0131 otomatik olarak tetiklerdi. Genellikle geli\u015ftirme ortam\u0131nda <code class=\"language-\">False<\/code> olarak ayarlan\u0131r.<\/li>\n<li><code class=\"language-\">tags<\/code>: DAG’lar\u0131 Airflow UI’da kategorize etmek ve filtrelemek i\u00e7in kullan\u0131lan etiketlerdir.<\/li>\n<\/ul>\n<\/li>\n<li>G\u00f6rev Tan\u0131mlar\u0131 (<code class=\"language-\">start_task<\/code>, <code class=\"language-\">hello_task<\/code>, <code class=\"language-\">world_task<\/code>, <code class=\"language-\">end_task<\/code>):<\/strong> Her bir g\u00f6rev, bir operat\u00f6r (\u00f6rne\u011fin <code class=\"language-\">BashOperator<\/code> veya <code class=\"language-\">PythonOperator<\/code>) kullan\u0131larak tan\u0131mlan\u0131r.\n<ul>\n<li><code class=\"language-\">task_id<\/code>: Her g\u00f6revin benzersiz tan\u0131mlay\u0131c\u0131s\u0131d\u0131r.<\/li>\n<li><code class=\"language-\">bash_command<\/code>: <code class=\"language-\">BashOperator<\/code> i\u00e7in \u00e7al\u0131\u015ft\u0131r\u0131lacak Bash komutunu belirtir.<\/li>\n<li><code class=\"language-\">python_callable<\/code>: <code class=\"language-\">PythonOperator<\/code> i\u00e7in \u00e7a\u011fr\u0131lacak Python fonksiyonunu belirtir.<\/li>\n<\/ul>\n<\/li>\n<li>G\u00f6rev Ba\u011f\u0131ml\u0131l\u0131klar\u0131 (<code class=\"language-\">start_task >> hello_task<\/code> vb.):<\/strong> <code class=\"language-\">>><\/code> ve <code class=\"language-\"><<<\/code> operat\u00f6rleri, g\u00f6revler aras\u0131ndaki s\u0131ray\u0131 ve ba\u011f\u0131ml\u0131l\u0131klar\u0131 tan\u0131mlamak i\u00e7in kullan\u0131l\u0131r. <code class=\"language-\">start_task >> hello_task<\/code> ifadesi, <code class=\"language-\">hello_task<\/code>'\u0131n ancak <code class=\"language-\">start_task<\/code> ba\u015far\u0131yla tamamland\u0131ktan sonra \u00e7al\u0131\u015fmas\u0131 gerekti\u011fini belirtir. Bu, i\u015f ak\u0131\u015f\u0131n\u0131z\u0131n mant\u0131ksal ak\u0131\u015f\u0131n\u0131 olu\u015fturur.<\/li>\n<\/ul>\nBu DAG'\u0131 kaydettikten sonra Airflow UI'ya giderek DAG'lar listesinde <code class=\"language-\">ilk_airflow_dag<\/code>'\u0131 g\u00f6rebilirsiniz. Orada DAG'\u0131 etkinle\u015ftirebilir ve manuel olarak tetikleyebilir veya zamanlay\u0131c\u0131n\u0131n devreye girmesini bekleyebilirsiniz. Graph View'da g\u00f6revlerinizin g\u00f6rsel temsilini ve aralar\u0131ndaki ba\u011f\u0131ml\u0131l\u0131klar\u0131 net bir \u015fekilde g\u00f6receksiniz. Loglara bakarak her bir g\u00f6revin \u00e7\u0131kt\u0131s\u0131n\u0131 da inceleyebilirsiniz. Bu basit \u00f6rnek, Airflow'un temel g\u00fcc\u00fcn\u00fc ve esnekli\u011fini g\u00f6stermektedir.<\/p>\n<div class=\"uzman-ipucu\">\n Uzman \u0130pucu: DAG dosyalar\u0131n\u0131z\u0131 geli\u015ftirirken, s\u00f6zdizimi hatalar\u0131n\u0131 erken yakalamak i\u00e7in yerel ortam\u0131n\u0131zda Python linters (flake8, black) kullan\u0131n. Ayr\u0131ca, DAG'lar\u0131n\u0131z\u0131 Airflow UI'da etkinle\u015ftirmeden \u00f6nce k\u00fc\u00e7\u00fck bir veri setiyle test etmeniz, \u00fcretim ortam\u0131nda ya\u015fanabilecek sorunlar\u0131 \u00f6nleyecektir.\n<\/div>\n<h2>Ger\u00e7ek D\u00fcnya Senaryosu: ETL S\u00fcrecini Airflow ile Y\u00f6netmek<\/h2>\nAirflow'un g\u00fcc\u00fc, \u00f6zellikle karma\u015f\u0131k ve ba\u011f\u0131ml\u0131 g\u00f6revler i\u00e7eren ger\u00e7ek d\u00fcnya veri m\u00fchendisli\u011fi senaryolar\u0131nda ortaya \u00e7\u0131kar. En yayg\u0131n kullan\u0131m alanlar\u0131ndan biri, farkl\u0131 veri kaynaklar\u0131ndan veri \u00e7ekip i\u015fleyerek bir veri ambar\u0131na y\u00fcklemeyi ama\u00e7layan ETL (Extract, Transform, Load) s\u00fcre\u00e7lerinin otomasyonudur. Bu b\u00f6l\u00fcmde, basitle\u015ftirilmi\u015f bir ETL boru hatt\u0131n\u0131 Airflow ile nas\u0131l y\u00f6netece\u011fimize dair bir vaka analizi sunaca\u011f\u0131z.<\/p>\nVaka Analizi Senaryosu:<\/strong> Bir e-ticaret \u015firketinin g\u00fcnl\u00fck sipari\u015f verilerini i\u015fledi\u011fini d\u00fc\u015f\u00fcnelim. Bu veriler, farkl\u0131 sistemlerden gelmektedir:<\/p>\n<ol>\n<li>Sipari\u015fler:<\/strong> Bir SQL veritaban\u0131ndan (\u00f6rne\u011fin PostgreSQL) \u00e7ekilir.<\/li>\n<li>M\u00fc\u015fteri Geri Bildirimleri:<\/strong> Bir API arac\u0131l\u0131\u011f\u0131yla harici bir hizmetten al\u0131n\u0131r.<\/li>\n<li>\u00dcr\u00fcn Fiyat G\u00fcncellemeleri:<\/strong> Bir CSV dosyas\u0131ndan okunur.<\/li>\n<\/ol>\nAmac\u0131m\u0131z, bu farkl\u0131 veri kaynaklar\u0131n\u0131 bir araya getirip d\u00f6n\u00fc\u015ft\u00fcrd\u00fckten sonra, analiz i\u00e7in bir veri ambar\u0131na (\u00f6rne\u011fin Snowflake veya Google BigQuery) y\u00fcklemektir. Bu s\u00fcrecin her g\u00fcn belirli bir saatte otomatik olarak \u00e7al\u0131\u015fmas\u0131 gerekmektedir.<\/p>\n<pre><code>\nfrom airflow import DAG\nfrom airflow.operators.bash import BashOperator\nfrom airflow.operators.python import PythonOperator\nfrom airflow.utils.dates import days_ago\nfrom datetime import timedelta\nimport pandas as pd\nimport requests\nimport csv\n\n# Varsay\u0131lan arg\u00fcmanlar\ndefault_args = {\n 'owner': 'data_team',\n 'start_date': days_ago(1),\n 'depends_on_past': False,\n 'email_on_failure': False,\n 'email_on_retry': False,\n 'retries': 2,\n 'retry_delay': timedelta(minutes=10),\n}\n\n# Python fonksiyonlar\u0131 tan\u0131mlayal\u0131m\ndef extract_orders_from_db(**kwargs):\n # Ger\u00e7ek senaryoda burada bir veritaban\u0131 ba\u011flant\u0131s\u0131 a\u00e7\u0131l\u0131r\n # ve SQL sorgusu ile sipari\u015f verileri \u00e7ekilir.\n # \u00d6rnek olmas\u0131 i\u00e7in sahte veri \u00fcretiyoruz.\n print(\"Veritaban\u0131ndan sipari\u015fler \u00e7ekiliyor...\")\n orders_data = [\n {'order_id': 1, 'customer_id': 101, 'amount': 150.00, 'status': 'completed'},\n {'order_id': 2, 'customer_id': 102, 'amount': 200.00, 'status': 'pending'},\n ]\n df_orders = pd.DataFrame(orders_data)\n kwargs['ti'].xcom_push(key='orders_df', value=df_orders.to_json())\n print(\"Sipari\u015f verileri ba\u015far\u0131yla \u00e7ekildi.\")\n\ndef extract_feedback_from_api(**kwargs):\n # Ger\u00e7ek senaryoda burada bir API \u00e7a\u011fr\u0131s\u0131 yap\u0131l\u0131r.\n # \u00d6rnek olmas\u0131 i\u00e7in sahte veri \u00fcretiyoruz.\n print(\"API'den m\u00fc\u015fteri geri bildirimleri \u00e7ekiliyor...\")\n # response = requests.get('https:\/\/api.example.com\/feedback')\n # feedback_data = response.json()\n feedback_data = [\n {'feedback_id': 1001, 'customer_id': 101, 'rating': 5, 'comment': 'Harika \u00fcr\u00fcn!'},\n {'feedback_id': 1002, 'customer_id': 102, 'rating': 3, 'comment': 'Teslimat gecikti.'},\n ]\n df_feedback = pd.DataFrame(feedback_data)\n kwargs['ti'].xcom_push(key='feedback_df', value=df_feedback.to_json())\n print(\"Geri bildirim verileri ba\u015far\u0131yla \u00e7ekildi.\")\n\ndef extract_product_prices_from_csv(**kwargs):\n # Ger\u00e7ek senaryoda burada bir dosya sisteminden CSV okunur.\n # \u00d6rnek olmas\u0131 i\u00e7in sahte bir CSV dosyas\u0131 olu\u015fturdu\u011fumuzu varsay\u0131yoruz.\n print(\"CSV'den \u00fcr\u00fcn fiyatlar\u0131 \u00e7ekiliyor...\")\n csv_data = \"\"\"product_id,price,last_updated\n1,25.50,2023-10-26\n2,12.00,2023-10-26\n\"\"\"\n # StringIO ile sanal dosya olu\u015fturma\n from io import StringIO\n df_prices = pd.read_csv(StringIO(csv_data))\n kwargs['ti'].xcom_push(key='prices_df', value=df_prices.to_json())\n print(\"\u00dcr\u00fcn fiyat verileri ba\u015far\u0131yla \u00e7ekildi.\")\n\ndef transform_data(**kwargs):\n ti = kwargs['ti']\n print(\"Veriler d\u00f6n\u00fc\u015ft\u00fcr\u00fcl\u00fcyor...\")\n\n # XCom'dan verileri \u00e7ekme\n df_orders = pd.read_json(ti.xcom_pull(task_ids='extract_orders'))\n df_feedback = pd.read_json(ti.xcom_pull(task_ids='extract_feedback'))\n df_prices = pd.read_json(ti.xcom_pull(task_ids='extract_prices'))\n\n # Basit birle\u015ftirme ve d\u00f6n\u00fc\u015f\u00fcm\n df_merged = pd.merge(df_orders, df_feedback, on='customer_id', how='left')\n df_merged = pd.merge(df_merged, df_prices, left_on='order_id', right_on='product_id', how='left')\n df_merged['total_amount_with_tax'] = df_merged['amount'] * 1.18 # %18 KDV ekleme\n df_merged = df_merged[['order_id', 'customer_id', 'amount', 'total_amount_with_tax', 'status', 'rating', 'comment', 'price']]\n\n print(\"Veri d\u00f6n\u00fc\u015f\u00fcm\u00fc tamamland\u0131.\")\n kwargs['ti'].xcom_push(key='transformed_data_df', value=df_merged.to_json())\n\ndef load_data_to_data_warehouse(**kwargs):\n ti = kwargs['ti']\n print(\"D\u00f6n\u00fc\u015ft\u00fcr\u00fclm\u00fc\u015f veriler veri ambar\u0131na y\u00fckleniyor...\")\n df_final = pd.read_json(ti.xcom_pull(task_ids='transform_data'))\n\n # Ger\u00e7ek senaryoda burada bir veri ambar\u0131 ba\u011flant\u0131s\u0131 kullan\u0131l\u0131r (Snowflake, BigQuery, Redshift vb.)\n # df_final.to_sql('daily_sales', con=data_warehouse_connection, if_exists='append', index=False)\n print(\"Y\u00fcklenecek ilk 5 sat\u0131r:\")\n print(df_final.head().to_string())\n print(\"Veriler veri ambar\u0131na ba\u015far\u0131yla y\u00fcklendi.\")\n\nwith DAG(\n dag_id='ecommerce_etl_pipeline',\n default_args=default_args,\n description='E-ticaret sipari\u015f verilerini i\u015fleyen ETL boru hatt\u0131.',\n schedule_interval=timedelta(days=1), # Her g\u00fcn \u00e7al\u0131\u015ft\u0131r\n catchup=False,\n tags=['etl', 'ecommerce', 'data_pipeline'],\n) as dag:\n # 1. A\u015fama: Veri \u00c7ekme (Extract)\n extract_orders = PythonOperator(\n task_id='extract_orders',\n python_callable=extract_orders_from_db,\n )\n\n extract_feedback = PythonOperator(\n task_id='extract_feedback',\n python_callable=extract_feedback_from_api,\n )\n\n extract_prices = PythonOperator(\n task_id='extract_prices',\n python_callable=extract_product_prices_from_csv,\n )\n\n # 2. A\u015fama: Veri D\u00f6n\u00fc\u015ft\u00fcrme (Transform)\n transform = PythonOperator(\n task_id='transform_data',\n python_callable=transform_data,\n )\n\n # 3. A\u015fama: Veri Y\u00fckleme (Load)\n load = PythonOperator(\n task_id='load_data_to_data_warehouse',\n python_callable=load_data_to_data_warehouse,\n )\n\n # G\u00f6rev Ba\u011f\u0131ml\u0131l\u0131klar\u0131\n # T\u00fcm \u00e7ekme g\u00f6revleri paralel \u00e7al\u0131\u015fabilir, ancak d\u00f6n\u00fc\u015f\u00fcmden \u00f6nce bitmeli\n [extract_orders, extract_feedback, extract_prices] >> transform\n transform >> load\n<\/code><\/pre>\nBu DAG, e-ticaret ETL s\u00fcrecini ad\u0131m ad\u0131m y\u00f6netir. \u0130\u015fte bu DAG'\u0131n temel \u00f6zellikleri ve Airflow'un bu senaryoda nas\u0131l bir de\u011fer katt\u0131\u011f\u0131:<\/p>\n<ul>\n<li>Paralel \u00c7al\u0131\u015ft\u0131rma:<\/strong> <code class=\"language-\">extract_orders<\/code>, <code class=\"language-\">extract_feedback<\/code> ve <code class=\"language-\">extract_prices<\/code> g\u00f6revleri birbirinden ba\u011f\u0131ms\u0131z oldu\u011fu i\u00e7in Airflow bunlar\u0131 paralel olarak \u00e7al\u0131\u015ft\u0131rabilir. Bu, toplam \u00e7al\u0131\u015fma s\u00fcresini \u00f6nemli \u00f6l\u00e7\u00fcde azalt\u0131r.<\/li>\n<li>Ba\u011f\u0131ml\u0131l\u0131k Y\u00f6netimi:<\/strong> <code class=\"language-\">transform<\/code> g\u00f6revi, t\u00fcm \u00e7ekme (extract) g\u00f6revlerinin ba\u015far\u0131yla tamamlanmas\u0131n\u0131 bekler. <code class=\"language-\">[extract_orders, extract_feedback, extract_prices] >> transform<\/code> sat\u0131r\u0131 bu ba\u011f\u0131ml\u0131l\u0131\u011f\u0131 a\u00e7\u0131k\u00e7a tan\u0131mlar. Benzer \u015fekilde, <code class=\"language-\">load<\/code> g\u00f6revi de <code class=\"language-\">transform<\/code> g\u00f6revinin bitmesini bekler. Bu sayede veri tutarl\u0131l\u0131\u011f\u0131 sa\u011flan\u0131r ve i\u015f ak\u0131\u015f\u0131 mant\u0131ksal bir s\u0131rayla ilerler.<\/li>\n<li>XCom Kullan\u0131m\u0131:<\/strong> Her <code class=\"language-\">extract<\/code> g\u00f6revi, \u00e7ekti\u011fi veriyi Pandas DataFrame format\u0131nda JSON'a d\u00f6n\u00fc\u015ft\u00fcrerek XCom (Cross-Communication) mekanizmas\u0131 arac\u0131l\u0131\u011f\u0131yla Airflow veritaban\u0131na kaydeder. <code class=\"language-\">transform_data<\/code> g\u00f6revi ise bu verileri XCom'dan \u00e7ekerek birle\u015ftirme ve d\u00f6n\u00fc\u015ft\u00fcrme i\u015flemlerini yapar. Bu, g\u00f6revler aras\u0131nda k\u00fc\u00e7\u00fck miktarda veri payla\u015f\u0131m\u0131 i\u00e7in olduk\u00e7a kullan\u0131\u015fl\u0131d\u0131r. B\u00fcy\u00fck veri setleri i\u00e7in genellikle ge\u00e7ici bir depolama alan\u0131 (S3, GCS) kullan\u0131l\u0131r ve XCom'da sadece dosya yollar\u0131 veya meta veriler payla\u015f\u0131l\u0131r.<\/li>\n<li>Hata Y\u00f6netimi ve Yeniden Denemeler:<\/strong> <code class=\"language-\">default_args<\/code> i\u00e7indeki <code class=\"language-\">retries=2<\/code> ve <code class=\"language-\">retry_delay=timedelta(minutes=10)<\/code> ayarlar\u0131 sayesinde, herhangi bir g\u00f6rev ba\u015far\u0131s\u0131z oldu\u011funda Airflow, g\u00f6revi 10 dakika arayla iki kez daha deneyecektir. Bu, ge\u00e7ici a\u011f sorunlar\u0131 veya veritaban\u0131 kesintileri gibi durumlarda boru hatt\u0131n\u0131n kendili\u011finden kurtulmas\u0131n\u0131 sa\u011flar.<\/li>\n<li>\u0130zleme ve G\u00f6zlemleme:<\/strong> Airflow UI sayesinde, bu ETL boru hatt\u0131n\u0131n her bir \u00e7al\u0131\u015ft\u0131rmas\u0131n\u0131, her bir g\u00f6revin durumunu (ba\u015far\u0131l\u0131, ba\u015far\u0131s\u0131z, \u00e7al\u0131\u015f\u0131yor), loglar\u0131n\u0131 ve \u00e7al\u0131\u015fma s\u00fcrelerini kolayca izleyebilirsiniz. Bir g\u00f6rev ba\u015far\u0131s\u0131z oldu\u011funda an\u0131nda bildirim alabilir ve sorunun k\u00f6k nedenini loglardan h\u0131zl\u0131ca bulabilirsiniz.<\/li>\n<\/ul>\nBu vaka analizi, Apache Airflow'un karma\u015f\u0131k veri i\u015fleme boru hatlar\u0131n\u0131 nas\u0131l basitle\u015ftirdi\u011fini, otomatikle\u015ftirdi\u011fini ve daha g\u00fcvenilir hale getirdi\u011fini g\u00f6stermektedir. Veri m\u00fchendisleri i\u00e7in, bu t\u00fcr bir orkestrasyon arac\u0131, veri ak\u0131\u015flar\u0131n\u0131n tutarl\u0131l\u0131\u011f\u0131n\u0131, zaman\u0131nda tamamlanmas\u0131n\u0131 ve genel veri kalitesini sa\u011flamak ad\u0131na vazge\u00e7ilmezdir.<\/p>\n<div class=\"onemli-not\">\n \u00d6nemli Not: Ger\u00e7ek ETL senaryolar\u0131nda, veritaban\u0131 ba\u011flant\u0131lar\u0131, API anahtarlar\u0131 gibi hassas bilgiler Airflow'un Connections veya Variables aray\u00fczleri arac\u0131l\u0131\u011f\u0131yla g\u00fcvenli bir \u015fekilde y\u00f6netilmelidir. Bu, kodunuzu daha temiz ve g\u00fcvenli hale getirir.\n<\/div>\n<h2>Airflow ile \u0130\u015f Ak\u0131\u015flar\u0131n\u0131z\u0131 Daha Verimli Hale Getirme \u0130pu\u00e7lar\u0131<\/h2>\nApache Airflow, temel i\u015f ak\u0131\u015f\u0131 orkestrasyonunun \u00f6tesinde, i\u015f ak\u0131\u015flar\u0131n\u0131z\u0131 daha sa\u011flam, verimli ve y\u00f6netilebilir k\u0131lmak i\u00e7in bir\u00e7ok geli\u015fmi\u015f \u00f6zellik sunar. \u0130\u015fte deneyimli kullan\u0131c\u0131lar i\u00e7in baz\u0131 ipu\u00e7lar\u0131 ve p\u00fcf noktalar\u0131:<\/p>\n<ol>\n<li>Idempotency (Tekrarlanabilirlik) Sa\u011flay\u0131n:<\/strong> G\u00f6revlerinizin idempotent olmas\u0131n\u0131 sa\u011flamak, veri boru hatlar\u0131n\u0131z\u0131n dayan\u0131kl\u0131l\u0131\u011f\u0131 i\u00e7in kritik \u00f6neme sahiptir. Idempotent bir g\u00f6rev, birden \u00e7ok kez \u00e7al\u0131\u015ft\u0131r\u0131lsa bile sistem \u00fczerinde ayn\u0131 etkiyi yarat\u0131r. \u00d6rne\u011fin, bir veri y\u00fckleme g\u00f6revi, hedef tabloda mevcut verileri silip yeniden y\u00fcklemeli veya sadece yeni\/g\u00fcncellenmi\u015f kay\u0131tlar\u0131 eklemelidir. Bu, ba\u015far\u0131s\u0131z bir g\u00f6revi g\u00fcvenle yeniden \u00e7al\u0131\u015ft\u0131rabilmenizi sa\u011flar.<\/li>\n<li>XCom'lar\u0131 Ak\u0131ll\u0131ca Kullan\u0131n (K\u00fc\u00e7\u00fck Veriler \u0130\u00e7in):<\/strong> XCom'lar (Cross-Communication), g\u00f6revler aras\u0131nda k\u00fc\u00e7\u00fck miktarlarda veri (\u00f6rne\u011fin, dosya yollar\u0131, ID'ler, durum bilgileri) payla\u015fmak i\u00e7in harikad\u0131r. Ancak, b\u00fcy\u00fck veri setlerini XCom'lar arac\u0131l\u0131\u011f\u0131yla ge\u00e7irmek veritaban\u0131 performans\u0131n\u0131 d\u00fc\u015f\u00fcrebilir ve Airflow'un veritaban\u0131n\u0131 \u015fi\u015firebilir. B\u00fcy\u00fck veriler i\u00e7in bunun yerine, ge\u00e7ici depolama alanlar\u0131 (Amazon S3, Google Cloud Storage, Azure Blob Storage) kullan\u0131n ve XCom'lar arac\u0131l\u0131\u011f\u0131yla sadece bu depolama alanlar\u0131ndaki dosya yollar\u0131n\u0131 veya i\u015faret\u00e7ileri ge\u00e7irin.<\/li>\n<li>Sens\u00f6rleri Etkin Kullan\u0131n:<\/strong> Sens\u00f6rler, belirli bir ko\u015fulun (\u00f6rne\u011fin, bir dosyan\u0131n S3'e y\u00fcklenmesi, bir veritaban\u0131 kayd\u0131n\u0131n olu\u015fmas\u0131, harici bir API'nin yan\u0131t vermesi) ger\u00e7ekle\u015fmesini bekleyen \u00f6zel operat\u00f6rlerdir. \u0130\u015f ak\u0131\u015flar\u0131n\u0131z\u0131n d\u0131\u015f sistemlere ba\u011f\u0131ml\u0131l\u0131klar\u0131n\u0131 y\u00f6netmek i\u00e7in sens\u00f6rleri kullanmak, gereksiz i\u015flemci d\u00f6ng\u00fclerini ve kaynak t\u00fcketimini \u00f6nler. \u00d6rne\u011fin, bir veri i\u015fleme g\u00f6revi ba\u015flamadan \u00f6nce, ilgili verinin bir FTP sunucusuna ula\u015f\u0131p ula\u015fmad\u0131\u011f\u0131n\u0131 kontrol eden bir <code class=\"language-\">FTPSensor<\/code> kullanabilirsiniz.<\/li>\n<li>G\u00f6rev Gruplar\u0131 (TaskGroups) ve Alt DAG'lar (SubDAGs):<\/strong>\n<ul>\n<li>TaskGroups:<\/strong> Karma\u015f\u0131k DAG'larda benzer g\u00f6revleri mant\u0131ksal olarak grupland\u0131rmak i\u00e7in TaskGroups kullan\u0131n. Bu, Airflow UI'da DAG'\u0131n\u0131z\u0131n g\u00f6rselle\u015ftirilmesini basitle\u015ftirir ve okunabilirli\u011fi art\u0131r\u0131r. TaskGroups, sadece g\u00f6rsel bir gruplamad\u0131r ve performans \u00fczerinde do\u011frudan bir etkisi yoktur.<\/li>\n<li>SubDAGs:<\/strong> Tekrar eden i\u015f ak\u0131\u015f\u0131 kal\u0131plar\u0131n\u0131 soyutlamak i\u00e7in SubDAG'lar kullan\u0131labilir. Ancak, SubDAG'lar\u0131n baz\u0131 performans ve y\u00f6netim zorluklar\u0131 vard\u0131r (kendi Scheduler'\u0131, Webserver'\u0131 ve Worker'\u0131 gibi davran\u0131r). Genellikle TaskGroups, \u00e7o\u011fu gruplama ihtiyac\u0131 i\u00e7in daha iyi bir \u00e7\u00f6z\u00fcmd\u00fcr. SubDAG kullanmadan \u00f6nce iki kez d\u00fc\u015f\u00fcn\u00fcn ve TaskGroups'un yeterli olup olmad\u0131\u011f\u0131n\u0131 de\u011ferlendirin.<\/li>\n<\/ul>\n<\/li>\n<li>Operat\u00f6rleri ve Kancalar\u0131 (Hooks) Ak\u0131ll\u0131ca Kullan\u0131n:<\/strong> Airflow, bir\u00e7ok pop\u00fcler hizmet (AWS, GCP, Azure, Spark, Hive, S3 vb.) i\u00e7in \u00f6nceden olu\u015fturulmu\u015f operat\u00f6rler ve kancalar sunar. Kancalar, harici sistemlerle ba\u011flant\u0131 kurmak i\u00e7in soyutlanm\u0131\u015f aray\u00fczlerdir. Bu operat\u00f6rleri ve kancalar\u0131 kullanarak kendi \u00f6zel kodunuzu yazma ihtiyac\u0131n\u0131 azalt\u0131r, kod tekrar\u0131n\u0131 \u00f6nler ve daha standart, bak\u0131m\u0131 kolay i\u015f ak\u0131\u015flar\u0131 olu\u015fturursunuz. Kendi \u00f6zel operat\u00f6rlerinizi ve kancalar\u0131n\u0131z\u0131 yazmak, \u00e7ok spesifik entegrasyonlar gerektiren durumlarda g\u00fc\u00e7l\u00fc bir se\u00e7enektir.<\/li>\n<li>Hata Y\u00f6netimi ve Bildirimler:<\/strong>\n<ul>\n<li><code class=\"language-\">default_args<\/code> i\u00e7indeki <code class=\"language-\">email_on_failure<\/code> ve <code class=\"language-\">on_failure_callback<\/code> gibi parametrelerle hata bildirimlerini yap\u0131land\u0131r\u0131n. Slack, PagerDuty gibi ara\u00e7larla entegrasyon i\u00e7in \u00f6zel callback fonksiyonlar\u0131 yazabilirsiniz.<\/li>\n<li><code class=\"language-\">sla_miss_callback<\/code> kullanarak SLA (Service Level Agreement) ihlallerini izleyin ve bildirim al\u0131n. Bu, i\u015f ak\u0131\u015flar\u0131n\u0131z\u0131n zaman\u0131nda tamamlanmas\u0131n\u0131 garanti etmenize yard\u0131mc\u0131 olur.<\/li>\n<\/ul>\n<\/li>\n<li>Ba\u011flant\u0131lar\u0131 ve De\u011fi\u015fkenleri Kullan\u0131n:<\/strong> Veritaban\u0131 kimlik bilgileri, API anahtarlar\u0131 gibi hassas bilgileri do\u011frudan DAG koduna yazmak yerine, Airflow UI'daki \"Admin -> Connections\" ve \"Admin -> Variables\" b\u00f6l\u00fcmlerini kullanarak y\u00f6netin. Bu, g\u00fcvenlik sa\u011flar, kodunuzu daha temiz tutar ve farkl\u0131 ortamlar (geli\u015ftirme, test, \u00fcretim) aras\u0131nda kolayca ge\u00e7i\u015f yapman\u0131z\u0131 sa\u011flar.<\/li>\n<li>DAG Dosyalar\u0131n\u0131 D\u00fczenli Tutun:<\/strong>\n<ul>\n<li>B\u00fcy\u00fck ve karma\u015f\u0131k DAG'lar\u0131 birden fazla Python dosyas\u0131na b\u00f6lerek veya yard\u0131mc\u0131 fonksiyonlar\/s\u0131n\u0131flar kullanarak d\u00fczenli tutun.<\/li>\n<li>DAG'lar\u0131n\u0131z i\u00e7in iyi bir isimlendirme kural\u0131 benimseyin ve <code class=\"language-\">tags<\/code> \u00f6zelli\u011fini kullanarak kategorize edin.<\/li>\n<li>Her DAG'a a\u00e7\u0131klay\u0131c\u0131 bir <code class=\"language-\">description<\/code> ekleyin.<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n<div class=\"uzman-ipucu\">\n Uzman \u0130pucu: Airflow'un <code class=\"language-\">template_fields<\/code> \u00f6zelli\u011fini kullanarak Jinja \u015fablonlar\u0131n\u0131 operat\u00f6r parametrelerinde dinamik de\u011ferler i\u00e7in kullan\u0131n. \u00d6rne\u011fin, <code class=\"language-\">ds<\/code> (data_interval_start) veya <code class=\"language-\">next_ds<\/code> (data_interval_end) gibi de\u011fi\u015fkenleri Bash komutlar\u0131n\u0131zda veya SQL sorgular\u0131n\u0131zda kullanabilirsiniz. Bu, DAG'lar\u0131n\u0131z\u0131 daha esnek ve dinamik hale getirir.\n<\/div>\nBu ipu\u00e7lar\u0131n\u0131 uygulayarak, Apache Airflow ile olu\u015fturdu\u011funuz veri boru hatlar\u0131n\u0131 sadece \u00e7al\u0131\u015f\u0131r hale getirmekle kalmayacak, ayn\u0131 zamanda onlar\u0131 daha sa\u011flam, \u00f6l\u00e7eklenebilir, bak\u0131m\u0131 kolay ve operasyonel olarak verimli hale getireceksiniz. Airflow'un sundu\u011fu geni\u015f \u00f6zellik yelpazesini ke\u015ffetmek ve projelerinize en uygun \u00e7\u00f6z\u00fcmleri bulmak i\u00e7in s\u00fcrekli deneme yapmaktan \u00e7ekinmeyin.<\/p>\n<h3>Apache Airflow'un Avantajlar\u0131 ve Dezavantajlar\u0131 Nelerdir?<\/h3>\nHer g\u00fc\u00e7l\u00fc ara\u00e7 gibi, Apache Airflow'un da kendine \u00f6zg\u00fc avantajlar\u0131 ve dezavantajlar\u0131 bulunmaktad\u0131r. Bir projede Airflow kullanmaya karar vermeden \u00f6nce bunlar\u0131 bilmek, do\u011fru teknoloji se\u00e7imini yapman\u0131za yard\u0131mc\u0131 olacakt\u0131r.<\/p>\n<h4>Avantajlar\u0131:<\/h4>\n<ul>\n<li>Pythonik Yakla\u015f\u0131m:<\/strong> \u0130\u015f ak\u0131\u015flar\u0131 tamamen Python koduyla tan\u0131mlan\u0131r. Bu, Python geli\u015ftiricileri i\u00e7in \u00f6\u011frenme e\u011frisini azalt\u0131r ve karma\u015f\u0131k mant\u0131klar\u0131 uygulamay\u0131 kolayla\u015ft\u0131r\u0131r. Ayr\u0131ca, mevcut Python k\u00fct\u00fcphaneleri ve ekosistemiyle kolayca entegre olabilir.<\/li>\n<li>Dinamik \u0130\u015f Ak\u0131\u015f\u0131 Tan\u0131mlama:<\/strong> DAG'lar Python kodu oldu\u011fu i\u00e7in, i\u015f ak\u0131\u015flar\u0131n\u0131 dinamik olarak olu\u015fturabilir, ko\u015fullara ba\u011fl\u0131 olarak g\u00f6revleri ekleyip \u00e7\u0131karabilir veya parametreleri de\u011fi\u015ftirebilirsiniz. Bu, statik yap\u0131land\u0131rma dosyalar\u0131na dayal\u0131 di\u011fer ara\u00e7lara g\u00f6re b\u00fcy\u00fck bir esneklik sa\u011flar.<\/li>\n<li>Zengin Kullan\u0131c\u0131 Aray\u00fcz\u00fc (UI):<\/strong> Airflow'un web tabanl\u0131 kullan\u0131c\u0131 aray\u00fcz\u00fc, DAG'lar\u0131 g\u00f6rselle\u015ftirmek, g\u00f6revlerin durumunu izlemek, loglar\u0131 g\u00f6r\u00fcnt\u00fclemek, ge\u00e7mi\u015f \u00e7al\u0131\u015ft\u0131rmalar\u0131 incelemek ve i\u015f ak\u0131\u015flar\u0131n\u0131 manuel olarak tetiklemek i\u00e7in g\u00fc\u00e7l\u00fc ara\u00e7lar sunar. Bu, operasyonel ekipler i\u00e7in b\u00fcy\u00fck bir kolayl\u0131kt\u0131r.<\/li>\n<li>Geni\u015f Operat\u00f6r ve Kanca Ekosistemi:<\/strong> Airflow, AWS, GCP, Azure, Spark, Hadoop, SQL veritabanlar\u0131 gibi bir\u00e7ok pop\u00fcler platform ve hizmetle entegrasyon i\u00e7in zengin bir operat\u00f6r ve kanca koleksiyonuna sahiptir. Bu, bir\u00e7ok yayg\u0131n g\u00f6revi h\u0131zl\u0131 ve standart bir \u015fekilde uygulaman\u0131z\u0131 sa\u011flar.<\/li>\n<li>\u00d6l\u00e7eklenebilirlik:<\/strong> CeleryExecutor veya KubernetesExecutor gibi da\u011f\u0131t\u0131k y\u00fcr\u00fct\u00fcc\u00fclerle birlikte kullan\u0131ld\u0131\u011f\u0131nda, Airflow binlerce g\u00f6revi paralel olarak \u00e7al\u0131\u015ft\u0131rabilir ve b\u00fcy\u00fck \u00f6l\u00e7ekli veri i\u015fleme boru hatlar\u0131n\u0131 y\u00f6netebilir.<\/li>\n<li>Sa\u011flam Ba\u011f\u0131ml\u0131l\u0131k Y\u00f6netimi:<\/strong> G\u00f6revler aras\u0131ndaki ba\u011f\u0131ml\u0131l\u0131klar a\u00e7\u0131k\u00e7a tan\u0131mlan\u0131r ve Airflow, bu ba\u011f\u0131ml\u0131l\u0131klar\u0131n do\u011fru s\u0131rada ve ko\u015fullarda \u00e7al\u0131\u015fmas\u0131n\u0131 garanti eder. Bu, karma\u015f\u0131k veri boru hatlar\u0131n\u0131n tutarl\u0131l\u0131\u011f\u0131n\u0131 sa\u011flar.<\/li>\n<li>Hata Y\u00f6netimi ve Yeniden Denemeler:<\/strong> G\u00f6revler i\u00e7in yeniden deneme politikalar\u0131, zaman a\u015f\u0131mlar\u0131 ve hata bildirim mekanizmalar\u0131 yerle\u015fik olarak bulunur. Bu, ge\u00e7ici hatalara kar\u015f\u0131 dayan\u0131kl\u0131l\u0131\u011f\u0131 art\u0131r\u0131r.<\/li>\n<\/ul>\n<h4>Dezavantajlar\u0131:<\/h4>\n<ul>\n<li>\u00d6\u011frenme E\u011frisi:<\/strong> Yeni ba\u015flayanlar i\u00e7in Airflow'un kavramlar\u0131 (DAG, operat\u00f6r, sens\u00f6r, XCom, y\u00fcr\u00fct\u00fcc\u00fc vb.) ve da\u011f\u0131t\u0131k mimarisi biraz karma\u015f\u0131k gelebilir. Kurulumu ve yap\u0131land\u0131rmas\u0131 da ilk ba\u015fta zorlay\u0131c\u0131 olabilir.<\/li>\n<li>Kaynak Yo\u011funlu\u011fu:<\/strong> Airflow'un Scheduler, Webserver ve veritaban\u0131 gibi bile\u015fenleri, \u00f6zellikle b\u00fcy\u00fck \u00f6l\u00e7ekli ve y\u00fcksek frekansl\u0131 DAG'lar i\u00e7in \u00f6nemli miktarda kaynak (CPU, RAM, disk I\/O) t\u00fcketebilir. K\u00fc\u00e7\u00fck, basit g\u00f6revler i\u00e7in bazen a\u015f\u0131r\u0131ya ka\u00e7an bir \u00e7\u00f6z\u00fcm olabilir.<\/li>\n<li>Ger\u00e7ek Zamanl\u0131 \u0130\u015fleme \u0130\u00e7in Uygun De\u011fil:<\/strong> Airflow, batch (toplu) i\u015fleme ve zamanlanm\u0131\u015f g\u00f6revler i\u00e7in tasarlanm\u0131\u015ft\u0131r. D\u00fc\u015f\u00fck gecikmeli, olay tabanl\u0131 veya ger\u00e7ek zamanl\u0131 veri ak\u0131\u015f\u0131 i\u015fleme (stream processing) senaryolar\u0131 i\u00e7in Kafka, Flink gibi farkl\u0131 ara\u00e7lar daha uygundur.<\/li>\n<li>Veritaban\u0131 Ba\u011f\u0131ml\u0131l\u0131\u011f\u0131:<\/strong> Airflow'un t\u00fcm meta verileri bir veritaban\u0131nda saklan\u0131r. Veritaban\u0131n\u0131n performans\u0131 ve eri\u015filebilirli\u011fi, t\u00fcm Airflow sisteminin performans\u0131 ve kararl\u0131l\u0131\u011f\u0131 \u00fczerinde do\u011frudan etkiye sahiptir. Veritaban\u0131 y\u00f6netimi ve \u00f6l\u00e7eklendirilmesi ek bir operasyonel y\u00fck getirebilir.<\/li>\n<li>Tek Bir Scheduler Bottleneck'i:<\/strong> Airflow 1.x s\u00fcr\u00fcmlerinde tek bir Scheduler vard\u0131 ve bu bir darbo\u011faz olabiliyordu. Airflow 2.x ile birden fazla Scheduler \u00e7al\u0131\u015ft\u0131rma yetene\u011fi gelse de, Scheduler'\u0131n kararl\u0131l\u0131\u011f\u0131 ve performans\u0131 hala kritik bir konudur.<\/li>\n<li>Karma\u015f\u0131k Ortam Yap\u0131land\u0131rmas\u0131:<\/strong> Da\u011f\u0131t\u0131k bir Airflow kurulumu (\u00f6rne\u011fin Celery veya Kubernetes ile) kurmak ve y\u00f6netmek, a\u011f yap\u0131land\u0131rmas\u0131, g\u00fcvenlik duvarlar\u0131, hizmet ke\u015ffi gibi ek karma\u015f\u0131kl\u0131klar i\u00e7erir.<\/li>\n<\/ul>\nBu avantajlar ve dezavantajlar g\u00f6z \u00f6n\u00fcne al\u0131nd\u0131\u011f\u0131nda, Apache Airflow, orta ve b\u00fcy\u00fck \u00f6l\u00e7ekli, zamanlanm\u0131\u015f, ba\u011f\u0131ml\u0131 ve karma\u015f\u0131k veri i\u015fleme i\u015f ak\u0131\u015flar\u0131 i\u00e7in m\u00fckemmel bir se\u00e7imdir. Ancak, \u00e7ok basit g\u00f6revler veya ger\u00e7ek zamanl\u0131 ihtiya\u00e7lar i\u00e7in daha hafif veya farkl\u0131 \u00e7\u00f6z\u00fcmler de\u011ferlendirilmelidir.<\/p>\n<h2>Sonu\u00e7: Veri Orkestrasyonunda Airflow'un Vazge\u00e7ilmez Yeri<\/h2>\nApache Airflow, modern veri m\u00fchendisli\u011fi ve analitik d\u00fcnyas\u0131nda veri ak\u0131\u015flar\u0131n\u0131 y\u00f6netmek i\u00e7in vazge\u00e7ilmez bir ara\u00e7 haline gelmi\u015ftir. Bu rehber boyunca g\u00f6rd\u00fc\u011f\u00fcm\u00fcz gibi, Airflow, Python tabanl\u0131 DAG'lar arac\u0131l\u0131\u011f\u0131yla i\u015f ak\u0131\u015flar\u0131n\u0131 programatik olarak tan\u0131mlama, zamanlama ve izleme yetene\u011fi sunar. Karma\u015f\u0131k ETL boru hatlar\u0131ndan raporlama s\u00fcre\u00e7lerine, makine \u00f6\u011frenimi model e\u011fitimlerinden sistem bak\u0131m\u0131 g\u00f6revlerine kadar geni\u015f bir yelpazede otomasyon sa\u011flar. Dinamik yap\u0131s\u0131, zengin operat\u00f6r ekosistemi ve g\u00fc\u00e7l\u00fc kullan\u0131c\u0131 aray\u00fcz\u00fc sayesinde, veri ekipleri manuel m\u00fcdahalelerin getirdi\u011fi hatalar\u0131 ve zaman kay\u0131plar\u0131n\u0131 en aza indirerek operasyonel verimlili\u011fi art\u0131rabilir.<\/p>\n\u00d6zellikle b\u00fcy\u00fck \u00f6l\u00e7ekli veri platformlar\u0131nda, farkl\u0131 sistemlerden gelen verilerin entegrasyonu, d\u00f6n\u00fc\u015f\u00fcm\u00fc ve hedeflenen depolama alanlar\u0131na y\u00fcklenmesi s\u00fcre\u00e7leri, Airflow gibi bir orkestrasyon arac\u0131 olmadan y\u00f6netilemez bir karma\u015f\u0131kl\u0131\u011fa ula\u015fabilir. Airflow'un sundu\u011fu ba\u011f\u0131ml\u0131l\u0131k y\u00f6netimi, hata kurtarma mekanizmalar\u0131 ve detayl\u0131 izleme yetenekleri, veri tutarl\u0131l\u0131\u011f\u0131n\u0131 ve boru hatlar\u0131n\u0131n g\u00fcvenilirli\u011fini garanti alt\u0131na al\u0131r. Her ne kadar bir \u00f6\u011frenme e\u011frisi ve kaynak gereksinimi olsa da, sundu\u011fu esneklik ve \u00f6l\u00e7eklenebilirlik, bu zorluklar\u0131n \u00fcstesinden gelmeye de\u011fer k\u0131lar.<\/p>\nSonu\u00e7 olarak, e\u011fer veri ak\u0131\u015flar\u0131n\u0131z\u0131n karma\u015f\u0131kl\u0131\u011f\u0131 art\u0131yor, ekipleriniz manuel s\u00fcre\u00e7lerle bo\u011fu\u015fuyor veya veri boru hatlar\u0131n\u0131z\u0131n g\u00f6r\u00fcn\u00fcrl\u00fc\u011f\u00fcn\u00fc ve g\u00fcvenilirli\u011fini art\u0131rmak istiyorsan\u0131z, Apache Airflow g\u00fc\u00e7l\u00fc bir \u00e7\u00f6z\u00fcm sunar. Veri odakl\u0131 stratejilerinizde Airflow'u benimsemek, sadece g\u00f6revleri otomatikle\u015ftirmekle kalmayacak, ayn\u0131 zamanda veri ekibinizin daha stratejik i\u015flere odaklanmas\u0131n\u0131 sa\u011flayarak i\u015fletmenize \u00f6nemli bir rekabet avantaj\u0131 kazand\u0131racakt\u0131r.<\/p>\n<h3>S\u0131k\u00e7a Sorulan Sorular (SSS)<\/h3>\nApache Airflow hakk\u0131nda s\u0131k\u00e7a sorulan baz\u0131 sorular ve yan\u0131tlar\u0131 a\u015fa\u011f\u0131dad\u0131r:<\/p>\n<ol>\n<li>Airflow ne t\u00fcr projeler i\u00e7in uygundur?<\/strong>\nAirflow, \u00f6zellikle zamanlanm\u0131\u015f (batch) ve ba\u011f\u0131ml\u0131 g\u00f6revlerin oldu\u011fu veri i\u015fleme boru hatlar\u0131, ETL\/ELT s\u00fcre\u00e7leri, veri ambar\u0131 g\u00fcncellemeleri, raporlama otomasyonlar\u0131, makine \u00f6\u011frenimi model e\u011fitim ve da\u011f\u0131t\u0131m i\u015f ak\u0131\u015flar\u0131 gibi senaryolar i\u00e7in \u00e7ok uygundur. Genellikle d\u00fc\u015f\u00fck gecikmeli veya ger\u00e7ek zamanl\u0131 veri ak\u0131\u015f\u0131 i\u015fleme i\u00e7in tasarlanmam\u0131\u015ft\u0131r.<\/p>\n<\/li>\n<li>Airflow'u kurmak zor mu?<\/strong>\nAirflow'un temel bir kurulumu (\u00f6rne\u011fin Docker Compose ile) nispeten kolayd\u0131r. Ancak, \u00fcretim ortam\u0131nda y\u00fcksek eri\u015filebilirlik, \u00f6l\u00e7eklenebilirlik ve g\u00fcvenlik gerektiren da\u011f\u0131t\u0131k bir kurulum (Kubernetes veya Celery ile) daha fazla yap\u0131land\u0131rma ve y\u00f6netim bilgisi gerektirebilir. Airflow'un resmi belgeleri ve topluluk kaynaklar\u0131 kurulum s\u00fcrecinde olduk\u00e7a yard\u0131mc\u0131d\u0131r.<\/p>\n<\/li>\n<li>Airflow yerine ba\u015fka alternatifler var m\u0131?<\/strong>\nEvet, piyasada AWS Step Functions, Google Cloud Composer (Airflow'un y\u00f6netilen bir s\u00fcr\u00fcm\u00fc), Azure Data Factory, Prefect, Dagster, Luigi gibi ba\u015fka i\u015f ak\u0131\u015f\u0131 orkestrasyon ara\u00e7lar\u0131 da bulunmaktad\u0131r. Her birinin kendine \u00f6zg\u00fc avantajlar\u0131 ve kullan\u0131m durumlar\u0131 vard\u0131r. Se\u00e7im, projenizin \u00f6zel gereksinimlerine, mevcut bulut altyap\u0131n\u0131za ve ekibinizin yetkinliklerine ba\u011fl\u0131d\u0131r.<\/p>\n<\/li>\n<li>Airflow'da hata ay\u0131klama (debugging) nas\u0131l yap\u0131l\u0131r?<\/strong>\nAirflow'da hata ay\u0131klama i\u00e7in birka\u00e7 y\u00f6ntem vard\u0131r:<\/p>\n<ul>\n<li>Airflow UI Loglar\u0131:<\/strong> En s\u0131k kullan\u0131lan y\u00f6ntemdir. Ba\u015far\u0131s\u0131z olan bir g\u00f6revin loglar\u0131n\u0131 UI \u00fczerinden inceleyerek hatan\u0131n nedenini bulabilirsiniz.<\/li>\n<li>Yerel Test:<\/strong> DAG kodunuzu Airflow ortam\u0131na da\u011f\u0131tmadan \u00f6nce Python beti\u011fi olarak \u00e7al\u0131\u015ft\u0131rarak veya <code class=\"language-\">airflow dags test<\/code> komutunu kullanarak yerel olarak test edebilirsiniz.<\/li>\n<li>Airflow CLI:<\/strong> <code class=\"language-\">airflow tasks test [dag_id] [task_id] [execution_date]<\/code> gibi komutlarla belirli bir g\u00f6revi belirli bir \u00e7al\u0131\u015ft\u0131rma tarihi i\u00e7in manuel olarak test edebilir ve \u00e7\u0131kt\u0131s\u0131n\u0131 g\u00f6zlemleyebilirsiniz.<\/li>\n<li>Python Debugger:<\/strong> Gerekirse, g\u00f6revlerinizi \u00e7al\u0131\u015ft\u0131ran Python koduna bir debugger ekleyerek ad\u0131m ad\u0131m ilerleyebilirsiniz.<\/li>\n<\/ul>\n<\/li>\n<li>Airflow'da veri g\u00fcvenli\u011fi nas\u0131l sa\u011flan\u0131r?<\/strong>\nAirflow'da veri g\u00fcvenli\u011fi i\u00e7in \u00e7e\u015fitli mekanizmalar mevcuttur:<\/p>\n<ul>\n<li>Ba\u011flant\u0131lar (Connections):<\/strong> Veritaban\u0131 kimlik bilgileri, API anahtarlar\u0131 gibi hassas bilgiler do\u011frudan koda yaz\u0131lmak yerine Airflow'un ba\u011flant\u0131lar aray\u00fcz\u00fcnde \u015fifrelenmi\u015f olarak saklan\u0131r.<\/li>\n<li>De\u011fi\u015fkenler (Variables):<\/strong> Genel yap\u0131land\u0131rma de\u011ferleri ve di\u011fer hassas olmayan veriler i\u00e7in de\u011fi\u015fkenler kullan\u0131labilir.<\/li>\n<li>RBAC (Role-Based Access Control):<\/strong> Airflow 2.x ve sonras\u0131, farkl\u0131 kullan\u0131c\u0131lar i\u00e7in farkl\u0131 eri\u015fim seviyeleri tan\u0131mlaman\u0131za olanak tan\u0131yan rol tabanl\u0131 eri\u015fim kontrol\u00fcn\u00fc destekler.<\/li>\n<li>Ortam De\u011fi\u015fkenleri:<\/strong> Hassas bilgiler ortam de\u011fi\u015fkenleri arac\u0131l\u0131\u011f\u0131yla da sa\u011flanabilir.<\/li>\n<li>Entegrasyonlar:<\/strong> Kubernetes Secrets veya HashiCorp Vault gibi harici s\u0131r y\u00f6netim sistemleriyle entegrasyonlar da m\u00fcmk\u00fcnd\u00fcr.<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n<\/body><\/p>\n","protected":false},"excerpt":{"rendered":"Apache Airflow: Veri Ak\u0131\u015flar\u0131n\u0131 Otomatikle\u015ftirmek \u0130\u00e7in Kapsaml\u0131 Rehber Veri ak\u0131\u015flar\u0131n\u0131z\u0131n karma\u015f\u0131kl\u0131\u011f\u0131yla bo\u011fu\u015fuyor, manuel m\u00fcdahalelerle zaman kaybediyor veya hata…","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"csco_page_header_type":"","csco_page_load_nextpost":"","csco_page_subscribe_form":"","csco_page_contact_form":"","footnotes":""},"categories":[874],"tags":[],"class_list":{"0":"post-35778","1":"post","2":"type-post","3":"status-publish","4":"format-standard","6":"category-server","7":"cs-entry","8":"cs-video-wrap"},"yoast_head":"\n<title>Apache Airflow Nedir ve Neden Veri Ak\u0131\u015flar\u0131n\u0131z \u0130\u00e7in Kritik Bir Ara\u00e7t\u0131r?<\/title>\n<meta name=\"description\" content=\"Veri ak\u0131\u015flar\u0131n\u0131z\u0131n karma\u015f\u0131kl\u0131\u011f\u0131yla bo\u011fu\u015fuyor, manuel m\u00fcdahalelerle zaman kaybediyor veya hata takibinde zorlan\u0131yor musunuz? Apache Airflow, veri m\u00fchendisli\u011fi s\u00fcre\u00e7lerinizi, ETL i\u015f ak\u0131\u015flar\u0131n\u0131z\u0131 ve di\u011fer zamanlanm\u0131\u015f g\u00f6revlerinizi Python koduyla programatik olarak y\u00f6netmenizi sa\u011flayan a\u00e7\u0131k kaynakl\u0131 bir platformdur. Bu rehber, Airflow'un temel prensiplerinden ba\u015flayarak, nas\u0131l \u00e7al\u0131\u015ft\u0131\u011f\u0131n\u0131, neden bu kadar pop\u00fcler oldu\u011funu ve kendi veri orkestrasyon \u00e7\u00f6z\u00fcmlerinizi nas\u0131l olu\u015fturaca\u011f\u0131n\u0131z\u0131 ad\u0131m ad\u0131m a\u00e7\u0131klayacak.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/\" \/>\n<meta property=\"og:locale\" content=\"tr_TR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Apache Airflow Nedir ve Neden Veri Ak\u0131\u015flar\u0131n\u0131z \u0130\u00e7in Kritik Bir Ara\u00e7t\u0131r?\" \/>\n<meta property=\"og:description\" content=\"Veri ak\u0131\u015flar\u0131n\u0131z\u0131n karma\u015f\u0131kl\u0131\u011f\u0131yla bo\u011fu\u015fuyor, manuel m\u00fcdahalelerle zaman kaybediyor veya hata takibinde zorlan\u0131yor musunuz? Apache Airflow, veri m\u00fchendisli\u011fi s\u00fcre\u00e7lerinizi, ETL i\u015f ak\u0131\u015flar\u0131n\u0131z\u0131 ve di\u011fer zamanlanm\u0131\u015f g\u00f6revlerinizi Python koduyla programatik olarak y\u00f6netmenizi sa\u011flayan a\u00e7\u0131k kaynakl\u0131 bir platformdur. Bu rehber, Airflow'un temel prensiplerinden ba\u015flayarak, nas\u0131l \u00e7al\u0131\u015ft\u0131\u011f\u0131n\u0131, neden bu kadar pop\u00fcler oldu\u011funu ve kendi veri orkestrasyon \u00e7\u00f6z\u00fcmlerinizi nas\u0131l olu\u015fturaca\u011f\u0131n\u0131z\u0131 ad\u0131m ad\u0131m a\u00e7\u0131klayacak.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/\" \/>\n<meta property=\"og:site_name\" content=\"Kodlar\u0131n Gizemli D\u00fcnyas\u0131\" \/>\n<meta property=\"article:published_time\" content=\"2025-12-03T17:25:55+00:00\" \/>\n<meta name=\"author\" content=\"Fatih Soysal\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Yazan:\" \/>\n\t<meta name=\"twitter:data1\" content=\"Fatih Soysal\" \/>\n\t<meta name=\"twitter:label2\" content=\"Tahmini okuma s\u00fcresi\" \/>\n\t<meta name=\"twitter:data2\" content=\"30 dakika\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/\"},\"author\":{\"name\":\"Fatih Soysal\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\"},\"headline\":\"Apache Airflow Nedir ve Neden Veri Ak\u0131\u015flar\u0131n\u0131z \u0130\u00e7in Kritik Bir Ara\u00e7t\u0131r?\",\"datePublished\":\"2025-12-03T17:25:55+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/\"},\"wordCount\":4870,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\"},\"articleSection\":[\"Server\"],\"inLanguage\":\"tr\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/#respond\"]}],\"copyrightYear\":\"2025\",\"copyrightHolder\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#organization\"}},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/\",\"name\":\"Apache Airflow Nedir ve Neden Veri Ak\u0131\u015flar\u0131n\u0131z \u0130\u00e7in Kritik Bir Ara\u00e7t\u0131r?\",\"isPartOf\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#website\"},\"datePublished\":\"2025-12-03T17:25:55+00:00\",\"description\":\"Veri ak\u0131\u015flar\u0131n\u0131z\u0131n karma\u015f\u0131kl\u0131\u011f\u0131yla bo\u011fu\u015fuyor, manuel m\u00fcdahalelerle zaman kaybediyor veya hata takibinde zorlan\u0131yor musunuz? Apache Airflow, veri m\u00fchendisli\u011fi s\u00fcre\u00e7lerinizi, ETL i\u015f ak\u0131\u015flar\u0131n\u0131z\u0131 ve di\u011fer zamanlanm\u0131\u015f g\u00f6revlerinizi Python koduyla programatik olarak y\u00f6netmenizi sa\u011flayan a\u00e7\u0131k kaynakl\u0131 bir platformdur. Bu rehber, Airflow'un temel prensiplerinden ba\u015flayarak, nas\u0131l \u00e7al\u0131\u015ft\u0131\u011f\u0131n\u0131, neden bu kadar pop\u00fcler oldu\u011funu ve kendi veri orkestrasyon \u00e7\u00f6z\u00fcmlerinizi nas\u0131l olu\u015fturaca\u011f\u0131n\u0131z\u0131 ad\u0131m ad\u0131m a\u00e7\u0131klayacak.\",\"breadcrumb\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/#breadcrumb\"},\"inLanguage\":\"tr\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Anasayfa\",\"item\":\"https:\/\/fatihsoysal.com\/blog\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Apache Airflow Nedir ve Neden Veri Ak\u0131\u015flar\u0131n\u0131z \u0130\u00e7in Kritik Bir Ara\u00e7t\u0131r?\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#website\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/\",\"name\":\"Fatihsoysal.com\",\"description\":\"Blog - Yaz\u0131l\u0131m D\u00fcnyas\u0131 Tecr\u00fcbelerim\",\"publisher\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/fatihsoysal.com\/blog\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"tr\"},{\"@type\":[\"Person\",\"Organization\"],\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\",\"name\":\"Fatih Soysal\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"tr\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png\",\"contentUrl\":\"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png\",\"width\":512,\"height\":512,\"caption\":\"Fatih Soysal\"},\"logo\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/\"},\"description\":\"Kullan\u0131m ve kodlama m\u00fckemmeliyetini odak alan uygulamalar olu\u015fturma deneyimine sahip, profesyonel olarak 15+ y\u0131l \u00fczeri deneyime sahip bir yaz\u0131l\u0131m m\u00fchendisi.\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/author\/fatihsoysal\/\"}]}<\/script>\n","yoast_head_json":{"title":"Apache Airflow Nedir ve Neden Veri Ak\u0131\u015flar\u0131n\u0131z \u0130\u00e7in Kritik Bir Ara\u00e7t\u0131r?","description":"Veri ak\u0131\u015flar\u0131n\u0131z\u0131n karma\u015f\u0131kl\u0131\u011f\u0131yla bo\u011fu\u015fuyor, manuel m\u00fcdahalelerle zaman kaybediyor veya hata takibinde zorlan\u0131yor musunuz? Apache Airflow, veri m\u00fchendisli\u011fi s\u00fcre\u00e7lerinizi, ETL i\u015f ak\u0131\u015flar\u0131n\u0131z\u0131 ve di\u011fer zamanlanm\u0131\u015f g\u00f6revlerinizi Python koduyla programatik olarak y\u00f6netmenizi sa\u011flayan a\u00e7\u0131k kaynakl\u0131 bir platformdur. Bu rehber, Airflow'un temel prensiplerinden ba\u015flayarak, nas\u0131l \u00e7al\u0131\u015ft\u0131\u011f\u0131n\u0131, neden bu kadar pop\u00fcler oldu\u011funu ve kendi veri orkestrasyon \u00e7\u00f6z\u00fcmlerinizi nas\u0131l olu\u015fturaca\u011f\u0131n\u0131z\u0131 ad\u0131m ad\u0131m a\u00e7\u0131klayacak.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/","og_locale":"tr_TR","og_type":"article","og_title":"Apache Airflow Nedir ve Neden Veri Ak\u0131\u015flar\u0131n\u0131z \u0130\u00e7in Kritik Bir Ara\u00e7t\u0131r?","og_description":"Veri ak\u0131\u015flar\u0131n\u0131z\u0131n karma\u015f\u0131kl\u0131\u011f\u0131yla bo\u011fu\u015fuyor, manuel m\u00fcdahalelerle zaman kaybediyor veya hata takibinde zorlan\u0131yor musunuz? Apache Airflow, veri m\u00fchendisli\u011fi s\u00fcre\u00e7lerinizi, ETL i\u015f ak\u0131\u015flar\u0131n\u0131z\u0131 ve di\u011fer zamanlanm\u0131\u015f g\u00f6revlerinizi Python koduyla programatik olarak y\u00f6netmenizi sa\u011flayan a\u00e7\u0131k kaynakl\u0131 bir platformdur. Bu rehber, Airflow'un temel prensiplerinden ba\u015flayarak, nas\u0131l \u00e7al\u0131\u015ft\u0131\u011f\u0131n\u0131, neden bu kadar pop\u00fcler oldu\u011funu ve kendi veri orkestrasyon \u00e7\u00f6z\u00fcmlerinizi nas\u0131l olu\u015fturaca\u011f\u0131n\u0131z\u0131 ad\u0131m ad\u0131m a\u00e7\u0131klayacak.","og_url":"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/","og_site_name":"Kodlar\u0131n Gizemli D\u00fcnyas\u0131","article_published_time":"2025-12-03T17:25:55+00:00","author":"Fatih Soysal","twitter_card":"summary_large_image","twitter_misc":{"Yazan:":"Fatih Soysal","Tahmini okuma s\u00fcresi":"30 dakika"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/#article","isPartOf":{"@id":"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/"},"author":{"name":"Fatih Soysal","@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1"},"headline":"Apache Airflow Nedir ve Neden Veri Ak\u0131\u015flar\u0131n\u0131z \u0130\u00e7in Kritik Bir Ara\u00e7t\u0131r?","datePublished":"2025-12-03T17:25:55+00:00","mainEntityOfPage":{"@id":"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/"},"wordCount":4870,"commentCount":0,"publisher":{"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1"},"articleSection":["Server"],"inLanguage":"tr","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/#respond"]}],"copyrightYear":"2025","copyrightHolder":{"@id":"https:\/\/fatihsoysal.com\/blog\/#organization"}},{"@type":"WebPage","@id":"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/","url":"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/","name":"Apache Airflow Nedir ve Neden Veri Ak\u0131\u015flar\u0131n\u0131z \u0130\u00e7in Kritik Bir Ara\u00e7t\u0131r?","isPartOf":{"@id":"https:\/\/fatihsoysal.com\/blog\/#website"},"datePublished":"2025-12-03T17:25:55+00:00","description":"Veri ak\u0131\u015flar\u0131n\u0131z\u0131n karma\u015f\u0131kl\u0131\u011f\u0131yla bo\u011fu\u015fuyor, manuel m\u00fcdahalelerle zaman kaybediyor veya hata takibinde zorlan\u0131yor musunuz? Apache Airflow, veri m\u00fchendisli\u011fi s\u00fcre\u00e7lerinizi, ETL i\u015f ak\u0131\u015flar\u0131n\u0131z\u0131 ve di\u011fer zamanlanm\u0131\u015f g\u00f6revlerinizi Python koduyla programatik olarak y\u00f6netmenizi sa\u011flayan a\u00e7\u0131k kaynakl\u0131 bir platformdur. Bu rehber, Airflow'un temel prensiplerinden ba\u015flayarak, nas\u0131l \u00e7al\u0131\u015ft\u0131\u011f\u0131n\u0131, neden bu kadar pop\u00fcler oldu\u011funu ve kendi veri orkestrasyon \u00e7\u00f6z\u00fcmlerinizi nas\u0131l olu\u015fturaca\u011f\u0131n\u0131z\u0131 ad\u0131m ad\u0131m a\u00e7\u0131klayacak.","breadcrumb":{"@id":"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/#breadcrumb"},"inLanguage":"tr","potentialAction":[{"@type":"ReadAction","target":["https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Anasayfa","item":"https:\/\/fatihsoysal.com\/blog\/"},{"@type":"ListItem","position":2,"name":"Apache Airflow Nedir ve Neden Veri Ak\u0131\u015flar\u0131n\u0131z \u0130\u00e7in Kritik Bir Ara\u00e7t\u0131r?"}]},{"@type":"WebSite","@id":"https:\/\/fatihsoysal.com\/blog\/#website","url":"https:\/\/fatihsoysal.com\/blog\/","name":"Fatihsoysal.com","description":"Blog - Yaz\u0131l\u0131m D\u00fcnyas\u0131 Tecr\u00fcbelerim","publisher":{"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/fatihsoysal.com\/blog\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"tr"},{"@type":["Person","Organization"],"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1","name":"Fatih Soysal","image":{"@type":"ImageObject","inLanguage":"tr","@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/","url":"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png","contentUrl":"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png","width":512,"height":512,"caption":"Fatih Soysal"},"logo":{"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/"},"description":"Kullan\u0131m ve kodlama m\u00fckemmeliyetini odak alan uygulamalar olu\u015fturma deneyimine sahip, profesyonel olarak 15+ y\u0131l \u00fczeri deneyime sahip bir yaz\u0131l\u0131m m\u00fchendisi.","url":"https:\/\/fatihsoysal.com\/blog\/author\/fatihsoysal\/"}]}},"yoast_meta":{"yoast_wpseo_title":"","yoast_wpseo_metadesc":"","yoast_wpseo_canonical":""},"amp_enabled":true,"_links":{"self":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts\/35778","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/comments?post=35778"}],"version-history":[{"count":0,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts\/35778\/revisions"}],"wp:attachment":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/media?parent=35778"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/categories?post=35778"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/tags?post=35778"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}

{"id":35778,"date":"2025-12-03T20:25:55","date_gmt":"2025-12-03T17:25:55","guid":{"rendered":"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/"},"modified":"2025-12-03T20:25:55","modified_gmt":"2025-12-03T17:25:55","slug":"apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir","status":"publish","type":"post","link":"https:\/\/fatihsoysal.com\/blog\/apache-airflow-nedir-ve-neden-veri-akislariniz-icin-kritik-bir-aractir\/","title":{"rendered":"Apache Airflow Nedir ve Neden Veri Ak\u0131\u015flar\u0131n\u0131z \u0130\u00e7in Kritik Bir Ara\u00e7t\u0131r?"},"content":{"rendered":"