Udemy
Turn what you know into an opportunity and reach millions around the world.
Learn More
Your cart is empty.
Keep shopping
Apache Flink: Build Real-Time Streaming Data Pipelines
New

Apache Flink: Build Real-Time Streaming Data Pipelines

Build Apache Flink streaming pipelines with DataStream API, Kafka, event time, windows, state, checkpoints, and outputs.
Created byACHRAF ER-RAYA
Last updated 10/2026
English

What you'll learn

  • Ingest low-latency, real-time data streams from Apache Kafka and CDC connectors using the production-ready Flink Source API.
  • Process stateful event streams using the Flink DataStream API with tumbling, sliding, and session event-time windows.
  • Configure event watermarks, custom late-data handling, and RocksDB state backends for scalable out-of-order stream processing.
  • Build fault-tolerant pipelines with transactional two-phase commit sinks, distributed checkpoints, and automated state recovery.

Course content

5 sections • 7 lectures • 1h 12m total length

  • Core Architecture & Pipeline Design: Building High-Throughput Flink Applications11:57
  • Core Architecture & Pipeline Design: Building High-Throughput (PART2)8:41

Requirements

  • Basic familiarity with Java or Object-Oriented Programming (OOP) concepts, along with fundamental understanding of command-line tools and software development basics. No prior experience with Apache Flink or real-time stream processing is required.

Description

This course contains the use of artificial intelligence.

Master real-time data processing with Apache Flink. Build scalable, fault-tolerant, high-throughput streaming pipelines from scratch using production standards.

Who this course is for:

  • Data Engineers looking to master real-time distributed stream processing.

  • Software Engineers transitioning into data architecture and pipeline development.

  • Backend Developers working with Apache Kafka, CDC, or event-driven systems.

  • System Architects designing enterprise real-time analytics platforms.

What you will learn:

  • Build end-to-end streaming data pipelines using Apache Flink and Java.

  • Ingest streaming datasets from Apache Kafka, CDC connectors, and enterprise sources.

  • Process stateful event streams using the Flink DataStream API.

  • Implement tumbling, sliding, and session event-time processing windows.

  • Configure event watermarks to manage delayed or out-of-order data streams.

  • Utilize state backends including RocksDB for high-volume enterprise state sizes.

  • Deliver processing results into transactional production sinks with zero data loss guarantees.

  • Recover processing pipelines seamlessly using distributed checkpoints and savepoints.

Requirements:

  • Basic knowledge of Java or Object-Oriented Programming principles.

  • Familiarity with basic command-line navigation and development environments.

  • Basic understanding of data engineering or database concepts.

Final Project Description: You will build an end-to-end real-time financial fraud detection pipeline. The pipeline ingests live transaction streams from Apache Kafka, processes key aggregations across custom rolling time windows, isolates suspicious transactions via side outputs, and sink results directly to a production database—fully equipped with automated state recovery checkpoints.

Who this course is for:

  • Data engineers, software developers, backend engineers, and system architects who want to master real-time distributed stream processing and build production-grade event-driven pipelines using Apache Flink and Java.

Report abuse