Watermark alignment is a relatively new feature in Apache Flink. It lets you cope with the problem of needing to temporally join streams with mismatched event frequencies, e.g., one stream’s updates are much more frequent than those of the stream(s) with which you need to join it. In this video we’ll break the feature down, and relate how it can help you better manage your Apache Flink integration.

A Data Streaming Platform (DSP) is composed of six major inter-related components, covered in detail in this video. Each component has evolved over time to servce specific needs in the data streaming platform, but together form a whole that is greater than its sum.

Connect ERP data from SAP S/4HANA and ECC with any cloud or on-prem system to power real-time applications and analytics with Confluent's fully managed data streaming platform directly integrated with SAP Datasphere.

In this video, Confluent and AWS address enterprises' challenges in deploying generative AI and how Confluent Cloud and Amazon Bedrock empower organizations to build scalable, AI-enabled applications. We'll explore how Confluent's comprehensive data streaming platform enables you to stream, connect, and govern data at scale, creating real-time, contextualized, and trustworthy applications that differentiate generative AI.

Transitioning to a set of microservices can seem daunting. This video breaks down how the Designing Event-Driven Microservices course can help you navigate that journey.

This video covers setting an idle timeout on a watermark generator when joining data in Apache Flink. This can be used when you have two streams, one that has frequent updates, and one that has infrequent updates, and you need to join data without waiting for a fresh watermark from the infrequent one.

Apache Kafka 3.8 is here! On behalf of the Kafka community, Danica Fine highlights key release updates, with KIPs from Kafka Core, Kafka Streams, and Kafka Connect. Kafka Core: Kafka Streams: Kafka Connect.

In this video, we'll look at some techniques for evolving events by analyzing a specific use case in a banking fraud detection system.

When developing or debugging a stream processing pipeline with Flink SQL, it’s common to inspect each processing step's output to ensure data is being transformed properly. However, comprehending the resulting data stream's structure, distribution, and characteristics entails executing multiple ad-hoc SQL queries, which can be time-consuming and tedious. Additionally, isolating specific subsets of the stream for analysis or debugging often involves even more queries, adding to the complexity and time required.