The Speedcraft Lab·1d agoNetflix Runs 30,000 Flink Jobs, and Open-Source Autoscaling Saved One of Its Teams $1.1M a YearThat team cut Flink compute 58%, and Netflix runs the scaler at 0.45 utilisation, not 0.7
Venugopal Dabbara·2d agoBeyond Micro-Batching: Exploring Real-Time Mode in DatabricksFrom Kafka transactions to operational decisions: a hands-on experiment using Databricks serverless Spark Declarative Pipelines, and…
Zeynepbkn·Oct 23Data Streaming: Gerçek Zamanlı Veri Akışının Gücü ve Kafka EkosistemiDebezium’dan Kafka’ya, Flink’ten Spark’a: Veri akışının mimarisini, temel kavramlarını ve Python uygulamalarını anlamak.
Realyze·Oct 21We gave an AI assistant Alibaba Cloud’s shopper-alert tutorial.A shopper puts something expensive in the cart. Then they go quiet.
Dilara Begüm Kırkoğlu·Oct 20Open-source vs. Fabric: building a real-time CDC-style lakehouse with open dataWhat it took to rebuild a Kafka / Spark / Druid / Airflow style platform in Microsoft Fabric — every Wikipedia edit, live, plus a bulk load…A response icon2A response icon2
Lakshman Gupta·Oct 18Streaming vs Batch & Exactly-Once SemanticsInvestment Bank Interview, Question 5 — a detailed model answer for Senior Data Engineer / Data Architect roles
Aliakbar Hosseinzadeh·Oct 13Kafka Streams: What It Actually Solves and When It Makes SenseUnderstanding KStream, KTable, stateful processing, exactly-once semantics, and when to choose Kafka Streams over Flink or Spark.A response icon1A response icon1
Niteesh K R·Oct 8Why Python Variables Aren’t Enough: Understanding Managed State in PyFlinkHow PyFlink remembers past events, maintains information across streams, and makes stateful real-time processing possible.
Sudarshan Shubakar·Oct 5Exactly-once delivery with Flink on a Kafka stream — Learnings in a labPart 2 of stream-lab. Same job, same jar, same 30,000 trades. One setting flipped, one Flink TaskManager stopped in the middle of a run…A response icon1A response icon1
